You are on page 1of 88

Annexes

213

Annexe A Article New results on the helium stars in the galactic center using BEAR spectro-imagery

215

216

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

217

218

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

219

220

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

221

222

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

223

224

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

225

226

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

227

228

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

229

230

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

231

232

ANNEXE A. ARTICLE NEW RESULTS ON THE HELIUM STARS

Annexe B Article New results on the Galactic Center Helium stars


Contribution pour la confrence Galactic Center workshop 2002, relative au travail sur les toiles hlium.

233

234 ANNEXE B. ARTICLE NEW RESULTS ON THE HELIUM STARS, UPDATED

235

Astron. Nachr./AN 324, No. S1, 1 5 (2003) / DOI theDOIprex/theDOIsufx

New results on the Galactic Center Helium stars


Thibaut Paumard1 , Jean-Pierre Maillard1 , and Susan Stolovy2
1 2

Institut dastrophysique de Paris (CNRS), 98b Bd. Arago, 75014 Paris, France SIRTF Science Center, CalTech, MS 220-6, Pasadena, CA 91125, USA

Received 15 November 2002, revised 30 November 2002, accepted 2 December 2002 Published online 3 December 2002 Key words infrared: stars, Galaxy: center, stars: early type, stars: Wolf-Rayet, instrumentation: spectrograph, techniques: radial velocities PACS 04A25 The cluster of helium stars around Sgr A has been re-observed with the BEAR spectro-imager on CFHT, in the 2.06 m helium line, at a spectral resolution of 52 km s and on a eld of 40 . This new analysis conrms and completes a previous study at a spectral resolution of 74 km s and on a smaller eld of 24 , corresponding to the central parsec (Paumard et al. 2001). Nineteen stars are conrmed as helium stars. These observations led to a clear differentiation between two groups of hot stars based on their emission linewidth, their magnitude and their positions relative to Sgr A . The rst class of 6 members is characterized by narrow-line proles (FWHM 200 km s ) and by their brightness. The other, fainter in K by an average of 2 mag, has a much broader emission component of width 1,000 km s . Several of the emission lines show a P Cygni prole. From these results, we propose that the narrow-line group is formed of stars in the LBV phase, while the broad-line group is formed of stars in or near the WR phase. The division into two groups is also shown by their spatial distribution, with the narrow-line stars in a compact central cluster (IRS 16) and the other group distributed at the periphery of the central cluster of hot stars. HST-NICMOS data in Pa (1.87 m) of the same eld reveal a similar association. The identication of the Pa counterpart to the He I stars provides an additional element to characterize the two groups. Bright Pa emitters are found generally associated with the narrow-line class stars while the weak Pa emitters are generally associated with the broad-line stars. A few particular cases are discussed. This conrms the different status of evolution of the two groups of massive, hot stars in the central cluster. As a by-product, about 20 additional candidate emission stars are detected in the central, high-resolution 19 eld from the NICMOS data.

1 Findings from BEAR 97 He I 2.06 m observation


With the BEAR spectro-imager, an imaging FTS (Maillard 2000) the central pc of the Galaxy was observed in 1997 at a spatial resolution of 0.5 and spectral resolution of 74 km s in the He I 2.058 m domain, covering a eld of 24 . The observation provided a homogeneous set of fully resolved line proles. The spectro-imaging data were associated with Adaptive Optics data from CFHT in the K band (Lai et al. 1997) to check the possible confusion of sources. That particular study of the helium emission-line stars in the central parsec of the Galactic Center was published in Paumard et al. (2001). The main results can be summarized as follows:
1. 16 fully resolved P Cygni emission line proles, cleaned of ISM emission, of purely stellar origin, were extracted. 2. they were found to divide into two distinct classes, with narrow (FWHM 200 km s ) and broad-line proles (FWHM 1,000 km s ). 3. a difference in K of 2 mag between the two classes was measured. 4. the spatial distribution of the two groups is different, with the narrow-line objects arranged in a central cluster, and the other class dispersed in a ring beyond a radius of 0.3 pc from SgrA .


From these ndings it was concluded that the group of narrow-line stars can be considered as formed of stars in the LBV phase, and the other one of stars at the WR stage.


c 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0004-6337/03/S101-0001 $ 17.50+.50/0

236 ANNEXE B. ARTICLE NEW RESULTS ON THE HELIUM STARS, UPDATED

T. Paumard, J.P. Maillard, and S. Stolovy: Results on the Galactic Center Helium stars

2 Observations
New BEAR data in He I 2.06 m were obtained in June 2000 at higher spectral resolution (52 km s ). The eld, composed of three overlapping circular sub-elds, was wider, and the signal-to-noise ratio higher by a factor of 1.6. Pa HST/NICMOS observations were taken in 1998 of the central parsec with Camera 1 (Stolovy et al.1999) and of the central 4 pc with Cameras 2 and 3 (Scoville et al. 2003). Dithered Images were taken in lters F187N centered on the 1.87 m Pa line and in F190N for the nearby continuum. By subtracting a suitably scaled F190N mosaic image from the F187N mosaic, a map of the stellar and interstellar Pa emission can be obtained. Figure 3 shows the central region of the composite Camera 2 and Camera 3 Pa image, for which the central has a spatial resolution of .

3 He I stars and Pa emission


With the new BEAR data, almost all the stars mentioned in Paumard et al. (2001) are conrmed, except the star numbered N6. The Pa data show a bright, very small ISM feature and no stellar counterpart to this point-like He I emission. Four new broad-line stars are added. Two were out of the previously studied eld, and the better signal-to-noise ratio is responsible for the other two new detections. The star B5 was associated with IRS 13E. Maillard et al. (2003) have shown that there was indeed two emission line stars in the IRS 13E complex, namely IRS 13E2 and IRS 13E4. The broad line clearly detected in He I belongs to E2 only from Fabry-Perot imaging associated with adaptive optics by Cl net et al. (2003). The e line proles and locations of the 19 stars are shown in Fig. 1 and Fig. 2a. Fig. 2b clearly conrms that the narrow-line stars are generally much brighter in than the broad-line stars. For B11, the signal-to-noise ratio is just sufcient to claim a detection, but not to derive reliable line parameters. Fig. 3a shows that the He I stars are associated with the Pa emission stars. The narrow-line stars W cm without taking into (circles) are coincident with bright Pa emitters (mean intensity 2 10 account IRS 34W, Table 1), whereas the broad-line stars correspond to fainter Pa emitters (mean intensity 0.97 10 W cm ). Other Pa emitters are present, which may be also associated with He I emission, but too faint to have been detected with BEAR. A source extraction with the StarFinder procedure (Diolaiti et al. 2000) gives 52 point-like emission features in the high-resolution central eld, of which 43 are emission line stars with a high degree of certainty. The 9 other need further observation to rule out the possibility that these point sources are compact ISM features or incomplete continuum subtraction of stars. However, this result (Fig. 3b) is generally consistent with an independent analysis made by Scoville et al. (2003).

Table 1 Physical properties of the helium stars: K magnitude, full width at zero intensity (FWZI) of the 2.06 m He I line (km s ), Pa line ux in units of 10 W cm , calibrated from Pa emission in AF (Najarro et al. 1994). ID 180 is from the photometric list of Ott et al. (1999), HeI N3 is from Paumard et al. (2001).

210 490 520 470 460 500

GE D B A9 HFC6@8

ID N1 N2 N3 N4 N5 N7

Name IRS 16NE IRS 16C IRS 16SW IRS 16NW IRS 33SE IRS 34W

m 8.8 9.6 9.4 9.9 9.9 11.8

FWZI

3.46 2.14 2.09 1.15 1.12 0.14

1,400 1,500 1,900 1,400 1,300 1,600 1,200 1,800 1,600 2,000 1,100 1,500 940

R P I A9 G (QC6@8

ID B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12 B13

Name ID 180 IRS 7E2 IRS 9W IRS 15SW IRS 13E2 IRS 7W AF AFNW HeIN3 BSD WC9 IRS 29N IRS 15NE IRS 16SE2

m 12.4 12.2 11.8 11.5 11.0 12.0 10.6 11.5 12.7 10.6 9.9 11.2 11.8

FWZI

0.40 0.28 0.83 0.76 2.27 0.44 4.40 1.05 0.14 0.37 0.07 0.37 1.15

" #

$" %#

('

   

  

1 0

" #

3 0

$" &#

53 640 )

237

Astron. Nachr./AN 324, No. S1 (2003)

4 Discussion
The central cluster of helium stars is conrmed with a total of 19 members currently identied. The various conclusions on the characteristics of these stars from the rst paper, reviewed in the Introduction are conrmed: division into two groups from their linewidths (Fig. 1 and Fig. 2b), from their brightness and from their location (Fig. 2a). The difference of brightness (Fig. 2b) presents few exceptions which were already noticed in the rst paper. One of the narrow-line stars (N7, IRS 34W) is weak and one of the broad-line stars (B11, IRS 29N) is brighter than the average of the other stars of the same group. From a long-term photometric study (Ott et al. 1999) IRS 34W is indicated as a variable star. It was weak at the time of our observations (Ott et al., private communication). It was proposed in Paumard et al. (2001) that the group of bright, helium stars was made of LBV-type stars. The high intensity of Pa (Table 1), the variability of IRS 34W conrm that these stars are hot, mass-losing stars, still rich in hydrogen. On the contrary, the weakness of the Pa emission combined with the very broad helium line are consistent with the other group being more evolved stars. A few sources are exceptions IRS 13E2, AF, IRS 16SE2 showing a broad-line He I prole, but strong Pa emission. This apparent anomaly could certainly be due to the fact that the Pa lter is not perfectly adapted to distinguish between rich and poor hydrogen emitters. Since the Pa line is blended with another signicative helium line, He I (4-3) at 1.869 m, the intensity detected by the F187N can remain strong even if the hydrogen emission is intrinsically weak. Already mentioned, the weakness of the K magnitude of IRS 34W (Table 1) is due to the star being in a phase of enhanced intrinsic extinction. Naturally, the measured Pa intensity is extremely weak, a factor 14 lower than the mean intensity. However, all these elements conrm the different status of evolution of the two groups of massive, hot stars in the central cluster. Assuming that all these stars were formed in the same star formation event, the differences in evolutionary state would come from the differences in their initial mass. The Pa data can help to address the question of whether the identication of emission line stars in the central region is complete or not. Possibly, about twenty new stars, associated with weak Pa emission are detected in the central parsec (Fig. 3b). With only this indication, it cannot be concluded that they are more WR candidates. A deep, spectroscopic analysis using adaptive optics in the K band is needed. Besides more WRs, some of them could be Be stars, or could belong to the old star population as symbiotic or Mira-type stars in a phase of emission. At any rate, these data represent a new element in the census of spectral type in the central parsecs to better constrain the peculiar star formation conditions in this region of the Milky Way.

References
Cl net, Y., et al., 2003, these proceedings e Diolaiti E., Bendinelli, O., Bonaccini, D., Close, L., Currie, D., & Parmeggiani, G., 2000, A&AS, 147, 335 Lai, O., V ran, J.P., Rigaut, F., Rouan, D., Gigan, P., Lacombe, F., Lena, P., Arsenault, R., Salmon, D.A., Thomas, e J., Crampton, D., Fletcher, J.M., Stilburn, J.R., Boyer, C., & Jagourel, P., 1997, In: Optical Telescopes of Today and Tomorrow, A.L. Ardeberg (ed), Proc. SPIE 2871, 859 Maillard, J.P., 2000, In: Imaging the Universe in 3 Dimensions, E. van Breughel & J. Bland-Hawthorn (eds), ASP Conf. Serie 195, 185 Maillard, J.P., Paumard, T., Stolovy, S.R., & Rigaut, F. 2003, these proceedings Najarro, F., Hillier, D.J., Kudrizli, R.P., Krabbe, A., Genzel, R., Lutz, D., Drapatz, S., & Geballe, T.R. 1994, A&A, 285, 573 Ott, T., Eckart, A., & Genzel, R. 1999, ApJ, 523, 248 Paumard, T., Maillard, J.P., Morris, M., & Rigaut, F. 2001, A&A, 366,466 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher, M., & Yusef-Zadeh, F. 2003, ApJ, submitted Stolovy, S.R., McCarthy, D.W., Melia, F., Rieke, G., Rieke, M.J., & Yusef-Zadeh, F., 1999, in The Central Parsecs of the Galaxy, ed. H. Falcke, A. Cotera, W.J. Duschl, F. Melia, M.J. Rieke, ASP Conf. Ser., 186, 39

238 ANNEXE B. ARTICLE NEW RESULTS ON THE HELIUM STARS, UPDATED

T. Paumard, J.P. Maillard, and S. Stolovy: Results on the Galactic Center Helium stars

N 1)

N 2)

N 3)

N 4)

N 5)

N 7)

B 1)

B 2)

B 3)

B 4)

B 5)

B 6)

B 7)

B 8)

B 9)

B10)

B11)

B12)

B13) Fig. 1 Upper row, the 2.06 m He I narrow-line proles; three lower rows, the broad-line proles, all on the same velocity range of -2,000 to +2,000 km s . The intensities are in W m cm (Continuum: a few 10 for N1N5, or 10 for the other stars). Simple line models are overplotted allowing to determine the full width at a few 10 zero intensity.

239

Astron. Nachr./AN 324, No. S1 (2003)

a)

b)

Fig. 2 a) Positioning of the helium stars, b) full width at zero intensity (FWZI) against K magnitude plot. The empty squares represent the broad-line stars, while the lled diamonds represent the narrow-line stars. On plot b), two regions can be seen: all the narrow-line stars, except IRS 34W that is variable (see text) , are grouped with a magnitude brighter than 10, and all the broad-line stars, except one, have a magnitude fainter thant 10.5 (mean value 12).

a)

b)

Fig. 3 a) Pa map of the central 2 pc, with the inner comprised of the high resolution Camera 2 data. Diamonds indicate the locations of the broad-line stars while the narrow-line stars are indicated by circles. All of these stars show emission in this Pa lter. b) All emission line star candidates, in the central region, marked by square boxes. The intensity scale is stretched to show fainter emission.

240 ANNEXE B. ARTICLE NEW RESULTS ON THE HELIUM STARS, UPDATED

Annexe C Article The nature of the Galactic Center source IRS 13 revealed by high spatial resolution in the infrared

241

242

ANNEXE C. ARTICLE THE NATURE OF IRS 13

c %cU yR Q#W!WQw0D}  mu j~ kj l k m qzjrr m 0qup CR $qq$UyC$'a ay8 0 06dyUC$'Cc$2yR qd0 yRy  0$$R ' 0CdyR q $ UC8R4IRC0 C C2'0$U$pyRy$ d0$ UaR$R R'$R 06 6 %R Uyy$R$CyC 06#8 2q$URUc8$y %yU 0 Uaq $$% 0q U' $$ 'Iy8R C0aR y 6y$6RCUa0 $ a$q | $ t$ UaR$R URU'yR Ryd 'Iyd $ 6 { mu zuzj yxjr u sr q om l kj U 4I# aR Q2w0`0wwIvtk`cpn0Wi '6IC6@  h %a%6XwI'C BF t CI'$8w h FA 3 g@ f f 3Aw 1 3 e d x x 3 x 5 3F B 9 9wA 79 G D FA G 9 V B V 7 v 7 GA F 5 BF V G9wA A V 7 v 53 wG F 5w 1w GA Ua%UpEwIEDCF %1 $WaCICtaEG'IE7aURC63IaS6ya!'%D 1B V 3 @B x 3 A w F F H A v 5 G x w A w F 5 @ x w e W e 1cv68$C%''cB $2$`aC'$w V gs 6 6W w%'F C%E5aIx Rw 0Ia'yc5$gD $auvCIp60A $'uaEGC8E7a'CA wF3 B 1G 9 @A 7G9 F B G F BA u V3 w w9 f 5 BF VG9w x w B V v B w w9 F9 w u w F F 3 f x wA 3 Aw B A 3 w 5 w F B x w x A @ w R'cFE1cBWE5U4URgdICuR'60)E56RIICE5iR80aG CA x w F w G 5 GB v w A V 3 A F 5 y%'zac$CdaCC$w @ %'G RacB EC6Rw %'o5 r w F w 9 w v 3 1 bV A 3 w F
e ht f r q i h g f e b Du`sDp0!Qd ca

b u 9A3F9 w wG AwF9 @ V3AF 5 @ w F wG 9 5 BF3@ BF w B xw B w yv06CUypp'U%E7V $60'$w %'I%aEGCg'y5 Ix'9y'A w F 5 G w D 3 D w F GF x V wA 3 @ B HV 3 79 7 5 7 g G V V9 7 G w %'i$A 6g%''dREw4C6E9%2Dp6%'%8IdyEGpEbUgaI5 G A A w 5 3 @ @B 'C6C% t aw C%aUEaSc'CRC8%cC''kW$'A t gu f @ A 7 G 9 wV v 5B 9 3 G F x w F 7 uB A F F 3 wB v 5 G w 5 3 @ 3AF @ w D9 w9 w 1 GAw x w BF 5 w B w D bF A V wF9 w H r 6C'RUk'RI'F aC)Rg C$8xkyU63EV'UkI20p )h 1 B x w 1 A V w F9 w F B v B x 5 7 GA A 79 w F 1 G Aw 9 w V 5 B9 B 1 IE7I463cVC0A C$E5WE58gI$'CI'%'saCXn%E5cB$cG''9w 1 G Aw 9 3 V 9 3 w D bF V 3 AF @ w D9 F w A w B x w G 9 F @ wT u G A V w F aCXYa2EVw t ayU60'Rg'y5 C8CRUy$63cVC9 1 G Aw 5 B9 B w w G 9 w B9 G D 1 @ V 3 7F @ 3 wA 3 f w aCXaEGUCE941 4R'cFU$aG tEbEV6g'a'6a8xB t P 6g'F 5 3 99 V FG 5 V CEw6IcFcB'9 3 yURg3 CcFE1cBV aEG'Ic7$UR'A iIa41 b u F9w u F xw B 5 BF V G9w 3 F3 w x 3 EbcVaCIa2I0'RU9 Ra0'%a%'qa'$XCI2H 0$IaaG @ V 3 A w 5 w v f 3 AF @ w D x w A w9 u G w F f wA Gw w A w 9 F 5 w 5 G D 1 V 3 A w w9 GF B x w V G9 wA 5 wFw G wA 3 9 @ A 7 G9 x w BF w B VA 3 w w 60aU'$E5Raua'R''X'aqRw C%aUR%'y5 IxvEb'a%'F x 5 3 A 3 w D D 3 9 @A 7 G9 A V wF9 F B 3w w 5 h w g66Rg%6Rw C%aUaE3EV'U$E56Df t I6vP p! X%x h H e w G w 9 3 @ F 9 w u w F 5B 5 GB F 7V G 9 w A V 3B F 3 D 9 3 F H x w v A w 1 ve I'aURg%'EdaE'IEaURCv6ECgU4qSuRa0w 93 Vw xw G @ B F w 5 B B V D 1 @ wAG 1 3 fQee ag x%R8x t 'A 9%'G acG'9yEwI4aG Ca26aa6W h F q p! XUF h q 2adaG 'RE'a6V gc%g0aC0RC'aG @ 9 3 9 w D @ 9 wV w F w v A 3 x 5B w u 9 A G F @ w A A 9@'Iadau'I6Iay5 $a%'tC06vP  y!EGa'y h BF D G w BF D 3 x 3 w G F w x 3 w F GF f h b V GF xA 3 G 5 G 9 3 w 1 3 W06$u ada0A 46@ q & r I'F aCXRa4E%'F w 1 G Aw 9 w v 3 1B w f h w 9 w 7 B 5 @ wF V g@ w D9 A G C BF t 6P p! X%x h R%$tI%CR'dEwyARg'avvP p!

 R R '0E R` UT '0%W6c P H F E C T R 8 T R T R T R YCUaRW0U)X9U UW0U UEa V2SR U U0y$ U0 $R4 Ua$RSR U6C$U$ 6Q6IGDB 06RC yk 0y $ y''aC0u 8R RIq$Uk y (' % $ "  8 R U RIk2p$ q6RR#&A@$)iR C0U)9C' $dR CCCa$u 8 R 0C y 6 UC0$U U$'$U$8a0y$0 U $ C$$' y Uc8a7U6up $ 0 6$U ay a C08ykc$RkU $$0C$ 0 Uy$R 0$ y 0y4RIRa d4qpIC '$y v 5 R U U0Cy0#yR UaC0a $y U'ac$Rg0C0aRdqp$4c0y&yk2 $UvR ' 0a0$Uk y6p 60 $ 2$U%R $yay$$ygR yyC0% CaC aCC4 $R $ (' % $ " Cy $ 484yy$ k 0$ Ua0C%0'U$ C$Uy$RU y$% $ %URg32C$$#)qp   ICU'1p$aC $$02'yRIUa 20CUaRk'$%0 ' 0$0$04  $8Ua 6uyR2C$$ $%y R0q c  0aR0U g 0 8I$q000ak  C$yy y 2$ CR $uR a R600aqCE yR2 $UCCCaRC084' yyCC0aR'y (' % $ "  URRu)&#iR UC8R$'CdRRCky0C$R0$v vCU'0!Rd0 0CU6 y Ry0'RCp60 U8ap) R2vkyRUd $ 2y 082yCU'y R 0C $ $0$  p2Uv8 #y 4R C0C0'$$4 80CCR2$ $yC$R$ $  U$8Iap0CI a   4 auyv y $dCCU'a pU $d0RCC0$ $U 8dR 6)  6d  'U'cak'0$ $ U UI 8ac6 a8 00 yR0 8$' y6kqpa00$R8R0 y aU' $tC$k CU' $RCCac6 a'yC$U% a 8I$ %u #q2yy$$%uR u $Uk 'y$ CCIk8 8%# y C0pC8U' yk yR RC$R4a$0ayyky4 $U q0R 20 UaR ay0y6yyU$y 0$ qy RC$R0$%I00a 6q2 U y$6CUaRC0 U'c 6 $2$U#q$R U U'$R UEa 0y8g $ R0 U8R 0)'U $C0Cpp RR0 0C0 a286RRRqya qa' y$RU%$ 0 a8R R6t%6yC$RuI6p Rk4%%'C I0C0C$0 C6p6 0$U$I0 u06%aRUa RyU$2p U #a2$ca$R I$ $p4 u8%Xv8vy#aXI  8uIqv48)))8tv u l l l h r h l r l ##R}t0 t#i0' h ~ | dCd#}{ zgw4ut2spqponkCiged# y x v j r m jl j h f

QeeWfeQ V7 aa6gIadcb%a`
243

AwF V x 53 b u xwFAw9 B w u t sr q i 0C3dg6%ydCU'%E5gvEVEVB %ph Y WV 7 S Q P 9 B H G 5 F B @9 7 5 3 1 (&  " !    6XIUT 8RCEAI%%28EDCA '%86420)'%$# 

tW r l u w tW T t t V w X R t tW y Vt t R R X R t t u w yt Wt t R y iSy5qi 1)2Fxiqi ixxt#2wUvqy1xi 2wxyA'Ut #2g{{cig d8{c#g UA#22t T# R T v w yy w t X R V y R R i r v RR t prW w T yt t t RW Wt w X R t W V pr r qXi#xjt#vudji2A%qynxUUsy#v#2tiUqRUeaxxGW na1Ffx'i1i2525FwixU'U#Fc}ca2rnp W T XtW R u W Rt t yt R wytW W T R ~ r p t wWy W Ry w y T w V R u w R x r l #xU5W nU1iix12UAV2wxnT2xUqRiGiUIi2i iUGq ncdj#2x)UU5)S2ie#Gi#TUuUqin)sG1) t R Xt VR t t r tR t y R yt Wtyt X W #iU1iU#%22wuU#UqR1CS12i 52ul ex1nxt 1ie2wxqyeG218Uy1#U5j8h X t tt X tW y t y V rs ~ r fd b ` Y t v v RW z u w x t y t R R y t v w y t W v t V r p R y tW T R y wW wWy t W R U #quwips#2it2aR UCex1)nG1G2xi# } ne'1#e2%Uxgnvn#2x)ix Uxy u w t w y r t R vt X u w Wt R V R t V w x RW V w w w R y T w T v V R T r l RW t y ty t V wW UA#2dcUxei%UxeAqyw2aR wpxaxgd2W u 2nx#2#2iUiyqyU1ed511#2#g r l u w yyt W v t u w W w t y w R y 1)kUF1#G2xijixU2GCiG2w#T#Gwv rs ~ r fd b ` Y t R w h X R v t tW V ~ r r eG21%8Uayxe#Gv{#Uy#Gw2xkU 2 r y Vt x RW t u w tW T wy t tW y X w r r t t w X R t RW V ~ r p ~ r r t u w W R W 2wqya5gdC#k25i2ixxt i #iqX2 u~ Gif#2#Us22#Ux}u nG 2#U2#2w ty w xW R v y w R V w R y p r l R R r r X R v z t y w V W t w x t i r X t v T t t G# xU#FkI)Gh222R v21de UvUt 8}U#fixxG2e#vv}1#qyi#iv X t X w x t i r V ty xy xW R v x RW X t vW wy v R t R s ~ r fd b ` Y t v v RW zh r l u w X R s r fd b ` 1nT2i%knyxU#exgdf1x2Uxi2#R 2aCeG218Uj1#UxC8U1)%U'#U5U2G1gejaY t y w y R X t T X t X X R r l u w V TW t h r l u w R t y X R s r fd b ` Y v V R x Gqy#GCGR e#xyn1nU1dFUix it 21)kUAxU1Ui5U2G18eaux#Uci t y x RW RW X R yW t X t ty tW w s p q q ph fd b ` ~ r fd b ` Y t R w ~ r fd b ` Y V h r l u w RW t Aaxgd5nU#FdxjixA1u1#wi eGGU8}geaY GG2age1xe#GviuG21gea#Tv{ap)kU51y Wt t w u w ytWt t XtWt W x X Rt R t t w RWt t l r Xt T X w tW tyW R RW t t u w T w 22FC2)1xGh#'1xG2xyi21WU#nqX#2qy#vi U5Gnx#i#2wxt 15U#c2xuixUf2weqR#i2 rytWt u w t v w R tW T wy t t R V w T W w W ty t w y tW wy wWt y t Rt V t 1)1xxGwUpn2Ux#2pqyixf2UAiiivxG2 UqRx)ixGnx2w qi2xGx%2VuU1yixI t V t y R W l yt yt v tyWR RW t t Gts2qRwdy# x2niqX11xUUxxGixix rsr q q p fd b ` Y XW R V T R h i2U#geF1x2AiU{QhF21r ytW T wy yt W v t w V R p q q p r ~ r 1x#2%)1u2#Ci#22v2G22U)a #2Uuae j 1tUx#2wuxi2jt 5iG# GnqX2xx2R X R y T tW X R tW T wy t t w tW T wy x RW T R R y w t y w r l t w Wt R r X R tW T wy t ytR XW w w 5i2yGx#2%U#GwxGh## 2wGy1) n%U#%qysU%qy#21f1tUx#2iF1en52u t R s q q p fd b ` Y w T o q q p fd b ` Y w R R h tW T w ty w t t V wW g r r X R p r r r X t w xt #hehU22Uvgea#ip%n22UvgeaU#UGU%{x#2y U22yix2#FG1UfF21%2s2a1U#ix y Vt x RW ttW X W v R u w tt X t 2wxqyeaxgdxyi1G2#U2w1ni%i X R v z t y t w Vw ttW XtWw tW xt r l #Uwixp5Gi22xi12#21') r t y r l W w g r R R X t V Ry t u w y R R W t V w w t V w 8UqR1A1y) 2#v#xe#yAU'#UynxUUfxGU#iftiA2R y t t R V t V t Wt w R tt X W t w X t v T s r q q ph fd b ` Y t RW R r p r X R r 12UAxqi Gt#2iUqy2w1xiwqyix}2t d 1iy#i#Si2288Ua%it i8tU5U#IUiF21#Udd21r rt w w tWR xt tw vR Xt X VtW y 1GnqX#2A2ixUix5# 22e21nAx2R t2i2RxFUxxeqR#G1U#ny#A2w u1#yqyUR W RWt T t uw X t R VW u tytW w r l R ytW T wy XtW R u w t x X Rt R tt v X R y y r l X R tW T wy X Wt u R R Xt w 1h)e5i212xW ixw22n21xW2AG#S2w 2R i2a)1t2dt 25iGA5WixuU{RCve#xw1U# R V w t X ty t t y Wt wW y r l t rs r fd b ` Y R h Vty xy i g t Xt R v e#Gin1ix#2#Vy2R 5# U1) 25au2angeca}Uk{nIvdv#A #Uxi2w X R r l y tW R V w t X R x t tW RW t t u w t R W t W R V w u X t RW t X R v U21dw#2w2xs8yh5G#2w ix1tU#vi 2w2UU5uFinU22AV}2xUqRv2xW d1x5Gxxu}Uz r u2r2p ixe5fexn2'iueAAqy#2A#Uw #wf21dUUAdxv##qyi#iei2U8#gekaY W x y R V r r r t RW R R T wy w V t w W t t r l u w R V yW t X t v T s r q q ph fd b ` t w V Wt wW R y w t w y t w XWR V TR Q r q uw RWR ty R p r X R r r l #ti 2wU2#duUiG#px #25GiG xUiUS r p iUp2we5Ux nF2aUd21) Vw w tW Twy t Xt wytW xt i rWt t w2hw hGx#2ixh1221pi%}uFAq2qRUR x2pxUGUp ia5s{z2aU5Gi22G2#v XtW w tW tWR s r qr h r uw y t w V w W t R W l Wt u R tW T wy V T T w X RW yt W v t #Udx2nwe5i2i#uGGywnGx)1xG2## xUG1pw 2wiT22u21iwxe Uariy22wnT21sUw R T t w wt X Wt uR ~ r q uw wytW u t v w XtW w tW xt  r q uw wytW R R hhx2ix5qi  22q iU2wiT21xAje}j VveGeG21#gecav2exF{xUi#22AVsn2xW u R R s ~ r fd b ` Y W R t h R R X R Q l V w t #fx A1UAeh1dUy2w1nAip1xGy U221C8UafdUGiqxA2U#i#I%xGU{Ru V V r X R R r l u w tt X t X t t s rh fd b ` Y u w x t W Ty Wt V w w t r W t tW s rh yy X R w R R V w X Wt t t iFU22a)xGw UwU#UA2ijxVuC#) t X T R V t w r X R r w R Tt z i#x #2UAf#221U ixd2#G15% z r V w tW T wy yt W v t y R XtW w tW y Wt VW u t i r r X }t#2 x#2S)1u2#vixk212c2R GwFi%UU1UR tWw V wy R t R Wt R x2dqyGwxnT2cU#ixdU#dU#dqyix ic t 21d2fxe#G1n2w #it 5#  Up 2uv w y r l y R XtR yt X W ut t tW  r r x R T r t x W ty t t tW w v R t v T w 2xU#xd}2u)sUqRx)jWixqyxiti2ii2ix 1e5U#e15i2}2SW2#juc12211)U#xe Xt RW R ty ytW T wy u w R R w yt wytW r l R yt R t x tWt X R u w t v T w y X R W v w w tW p r l X s q r fd b ` Y w lh X R v z t R T w W R 2u)kuis#UCX#2#UcG2#snSxt F21) i#FUG21geaf2VugfUeih2wxexinsUiT X R r r l r t x u w W t V t V t t t v T w x v wy tW X w ty W R v Ty R tW yW r V X w x t UsG1f)2u)8%Udx#T#iqXi2#y u#GwxnT2x#215U##Ce2w xUp#22v X R r l R X t X w yW Ry t X R wW v Wt ty R tW tty t X X Ty R y tytW V ~ w r V w u T W Uf21dAe#x1nT2t 5Ux#gn2Gxi#Uw 21#Un##SxGxiev2xW w2wiivx t W u XtW Ty Rt V t X T R V t w R T t z R X R #C2w p1xiGAn#2UA#2#uxshUz qynC2x#F#iadc2w dxxUqRiqU#eU#21i2aA#U X x Wt t t r l W u W T WR ty T Rt v r X R %5i2FuU{R%#jG1h)x GU#2wxyt xU# 2v15i2}CaC#U%x2s#AS5i2w)ui#T%e5W tW T wy Wt u t r l r t y V wWW R r ytW T wy V qr t u w tW w t ytW T wy x y w V t R y tW T wy V t Wt w t R X R v z t R R ty tW R tWt t w s rh fd b ` Y t t Wt t x 15iGiTiixU#'2#x#Uhix'k#UAV U#xUCix2iF2221}geaGts#uFq2qRUR nu2U52apn1GU1t xUx)qyxx2xU5qyU2UR #e2xiUc15iGdix2w2wx#xyije2W 12t d rp | xv tRWtR WttR R Wt RWR w R tW R tW t R x Wt t uw ytW T wy t w yy T X R V w uWt w xt R yW Ry t X R wW v Wt w t V w u X tW Ty R t V y ixwe5Ux#gn2G#wixUiGW 1#21w2R U221s5nUntd{AUU2G1#geS12t%8un# rs r yW t X R l h X R s r fd b ` Y t t h z rW Ry V t t y r l R y tW T wy t u w t gkU5)iTiixC2R 21)e1x#2hixvUi2w G#xUit2aR xU5W ng2#2dex#12x%qysUy y t XtW R u W Rt R R tWt X tW w tW w 5qi x uunqXxxh2R x)n##vxU5)iTi G2xisknGw xxU1I1Ut#UiG2q#u)scqy1Gs52nUct t w xW w y ty t t i ryW Ry V t W v R y tW W Rt r V R t R x R W l u w y u w Wty RW t t t v RW R V w yW R T w r t R wW R Xt R r l Xt V R R W w UCx)TS2xxGixf'xuqX #t#UxU2' xU5)y ii2y  ijxeAVne#iU1Uiwai1A)1UjxUq2x Wt w X yW Ry t X R wW v u w yy t w t t v w t Wt R X X Ty ytW T wy t w V ixUy'inTxU5)higuGG#sUw2qR#yxp#2wv qyik1uGixkUkqyix1tn5i2F#diGs tW T wy t t R V s y V t | o gh wW t 5i2##uUA5ueauyc8t#2## r w W w yW Ry y tW tty u w wyt 1Uis2F5U5)t2iCxt #SU2wxnT2xW y Vt Xt Wty v w t uw t y t V T WR R W T R t R ytW T wy y Vt t Wt t tt w Wt XW 2wxqy12#2p#UxuqX #w2R G#2xUUAV UqRi2#Uv#kec15i22wxyAvixki# 51i5Cn52w t i r t Ry t R W Ry V t R t v T w y W Ry t x  ~ r q u w wytW R y R t t V Ry t u #vUGUx#jeU5)iTihApqX#2iUxAu) } Uariys2p2wxnT2xUqRe' }qX#AUiUw R uw X Rty r l VW u y R Xt t ytW T w X R v z t yt R s h y w t R X R w XtWR V w svkh%Uy21#21dh#A2w 2'1 xGnqX1x#2y #Uw#ps122AVsC{wxnG2xn2i2hxU2 t t t u w w R X t X w s r q q ph fd b ` Y XW R V T R tW R R X t tW utW t i r t X t wW R Ty W u xWixyiCUyt#2hU#1nT2iGU8gea5UA#USQ xt 5ei##2x2w xi%}qX#1nX hfxC2w Gt RR V w u wytW RWt y R wW u Ty w wytW Wt w r t Vt wW R #xe#XvhtixGW 2wiT21xU512ieht#Uxi idUi%qy2wiT21xvqyixG2t dnG2exiAVsGU # t ~ r p t t w RW t t V w u Wt w y R y tytW tW  r q | r r h Rt z R TR o uw #VUUGq nwcdixuU2wx2xxu#GW v2t d 2v5Gxixh5qi qs U~ iy}2t 8{h1jA##UyUh2wxnT r tW T w g t R s Rh x Q u w RW t y X R v wy tW X t t ty t R x wy wW t y X tW R u W R t R i15i2yfddixX#2cjvkAU8#2uFUj51w#2#gjz 21AxVg2ijixewuGxGx1w1U5W ngU1iU5 wyt t t v Vt tW wWy R yW Rt v V TWt v w w yy w yt t R V t V TWy i rs q q q p XW R o 2wiT21xWg wv#wx#UqR#y1#2x)j5U1#1y i2wxdt#vu% F12UAsG##vy#vi222UhxUqRUp8h r l t Xt X w s ~ rh d b `Y t R w y Wt Wtt V wWt l VW uy RW i W W T w g R R 1)#1i#T#2eGG1)Ff8ejae2vqyx2R 2x1nmx2w #U5kjtxi2#h#GUAVfUedFv'x U5Uxx511ixVG#8{csy1iivxUm1rneR ##GUAVgUx11t2xW vi2ASi2Ugigeca5UA#USQ W R RW t y ty t wW g r t w X t T W R V r p R wW t y yy y V w u t X R V sr q q ph fd b ` Y XW R V T R P D (&4 E E D B@ 9 76 43 0 (& $ "     1( )G0 1I5)H GFCA%8%5)21)'%# ! 
244

ANNEXE C. ARTICLE THE NATURE OF IRS 13

x V 7 A 7 R 0"1yd0Xyevd& 1 1y""`a10R"UBFX1 1 `ye 7 x y x y j9 $ "c141U10RRU Tv# T`4Fva"e4 V x x rA PH G F E % 1Rf"ctS`TQghaye`e 1 XoXey`X0`Xe"exe 0RFRy`FyXeycw"cf`vTS1`"1us1"Xey`1 q c RR t er q 9 7 V U R R PH G F E % f yc"ex ip"RXv"yv"j@8r`$F4I`TUQIhEgA f1R0XR tee " " 4 da`Fj xes4TTTSzQIbSa01u4jtA 6 V c R R PH G F E 6 X 9 V PH G F E A 9 7 6 4 2 RR1s"Xre` Y""UTsW4U&RTRSQI3@eDCBj@8F 531 ) % 0@RjzFRt('R"R 1yyv"&1"e1ux $   #""!  t`at FpX"eoe01ua`RX`BRefX"pyv"410 e@0y1"`"ex`Rl"1@"X1XRF"ct0y`yvFwf1ux Xx$X`yeXpuj(""XeyU1wXe"`4v`BpEyd01x 1xpxe0f1R0XyFe411ys4v`j01xtUx`1yr X0Ut"xw ttl"dH@t@"xpXufX0"fx woftxo 

ce0(XxyH"d4H 4F`1f1seay(f"Roe 4 `yr1XRRw`R"yea` q Xyexycz011@XByXt`1XR14f('`x0"ReXs0Rxry10 y(p c(x4(U 1Rey$FF0pctft4v(F0H" 4sy(F0H y y 1yx1e@XE"R"e 4 0@j(f1z0yyF f{1f1yFx tyv0 1@ F  ctpU p(tdF0(f yUz0ya(vy 4 Xejf0R1yyUt c0FFy1Reasz0R1v ey1FR"fy10 F0(j 70lUU07(t`f eff0 zuF z4 E pjU0X} xX{ q w $ c @ d FF( v v R p w 4 t F ~ vf |( ~ V R R R A PH G F E V R R R A X 1"1`ydu4TTT(QIzdvX1RUBxewTT(p10XdfB0fyR q V sRez4v`wys`yf1r"Xy`R"zyv"s"R"ey%f(t` t x % q q A &x@xR`Fe@teySy d&xexye`X1ucveyyF30y c R R U R R q X 1y`dR"r"c``TvStx`sexez`Ttx`sEfFy q V r y c U j r r U y r UA yE4v`j0"cF4T`TIbxeevT`x0s@e411R"yF ay4FXeBsFxx`"Xjy`ye1"xFBrg4eFf@11y 9 q pe@BpcRsj"R@y$8yFjaR1aeBB`yt1c04ex` X`yX4ey"R"XURR"1ww0xyEweTTvu"e q r r r 7 f1TEfv`SET $ vSEIe`ySE7I1S1Eg`eyy x(tU xR 7 U c 7 xe 7 j x 7 r  q pn m s""lQo@t f 7 q % 1"Xof$I"da$fe"z"RX0`Xe"1dRxeBi r U x x c "RXj01e@Xj@`yxe1Xy1f"c TIX`"e TSX`x XFyfRy@a s9 $ l0y`R1@"Ru j(Be("aU 7 c x r x i R "RwH TSX`yX`k1y1eBf"R1xed TSX`x j x 7 7 9 RfI1X`1yc0XR1v & 1 `ye j9 $ "RB1e "R 7 xr xr 1"i 8 Xp1e@Xy10Eez1ey1ea""c 41U t e V 10RRU"1axeX@X"1!"e@XFyfRyX"eoA 9 i V c R R PH G F E A xe" w"RxelBw`yXfTTScQIe!01uyrtye h Y X 9 q s( 2 `yxe y(`j x`Tdg`yy$f0`Xe41 q xxxr t yEyv"E"c0"TTeX`1!oyF$Xy1X1B$`yz1o1FX$ Ea0Rxx1`R1Xyvyy(x1 s"R1jRe"yEq 7 hP S sQ s b gu W YT Y W Y q u F"1XuxeXQ"1sXP"Rya"$yT yv"y1s`XrB`$ yFUeF@0RyF 0d 0 @e a"pp41vxe `r `de s9 $ sF"l0R"Ew1e vadeBz`ywe$0Ryex"1yRUB"RSye4(r1Xy1@we1XuRvads R x 7 q s d y Q YT u QPg d b s dgT SQ Q YT W Pg dTP Q WPP d g u S WT C V x x rA PH G F E b u b s d S WT b Q Q Q YT u W Wg u Q YT g P S WT d s b W s W pXxe" p1"x"c4S"TgQIb` E X0xXyy0 ``Fx1"1XuR`"@`yxeXS1XR1@`yrFusx1XuRvaXW`XgXQ"RP r r P u b P d sQ q P d v Qg uP dg r Q YT u S WT dTQg bgQT W Q YT g yeyv"R0`$o1xFy0y tee "" 4 "R&1euTR Ff$FwXsEd1"1y"1euytsf`ye1XuRv4y`yRUSf"Rd1u x r x x x r q P Q S Q q QP S u T Q W V P d Y Qg T gTP W Y n h m k ih QP u b g u 1yx"TTe1R"eF1yFucX`XaUsrz&"R 0x`F0R"1p0RXg`f1"R"R"yRfXP"co`0lj(@RFf1 q q C Y W Y q P Qg uP s d W W V W Q YT u Qg T d S Q YT V Q W s b y d0`Xey`1@yEyv"&"1"t"coRU&a"`yEe ya"$f0y"1RcexaVUXxXSe"Rdey`yv"`ypU0"S1Xd"Q x r y s Q q W V QTg u b Qg sP W Qg b S WT WgT W V g Q S Q s dg 1Hy1"p`yey T1ye0 uxR"XRX`U0yf01 XsEEaPyFxR@R"Fuy`EB`7Fuy"yRaP@FR`feyyT RxF1yFuc5"XaUfc1" 531 "XRa"s`yX r `x 0f$eadW"fr`yEvxRwv`RtXsXs$rpUfe"caP`RXS@0RURP q 4 2 x Q bP g s Q b Q YT Td YT qu YP W q Q i hgQ bd b W YT W VQT SQ A I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$" 
245

q eB1XRv4F0R""Rzy d0Rwyx1""RzXBx`1xfzet 1y`r0xXyy0oe"wXwFyv"zy10`e$"RR0y`FeaR`Xx"c   &(#v&{I#ws"d@xyf q 1yf`X1Xyaeed1$bt "1 q 0j("Rz4eycexa`"X`R$e 4a`da"X`Xyy XteaRFRR0EFco1e@XpT 1 `XR`R0Rr`$"4Tr 1`X0`t A q  V x R 76 U c 7 q FXRey1TEIdv`SEUef`1XRy1&RdgU`RBFzye4( q 7 "X1XyaeRU`azEfe"sa`y$F`"U1Xxez"Xf"R q 6 `1Xe1XR"XRt R0`t 1`Rd@ye4(x1uWxe1Xyy1  q e&v 1 y d84e4(jRXyee1$ae"`XR0Rj1XReFd`e04e" 1yRUE`yrX4ey`yUeaexa0yweXXRx1XReF`r"11XR FfzRefxy%X`XdR41BX&1TSExe&v`ScEx x R 7 U c 7 q 1y4FrXeys"70x@0R"z $ 4eyj1`z`yoezy"`RF  q v`ypp1XRRalT 1 1yp1"p4a`dFRRa 1"aR0Rjeay(p1Xy RFfaefyee@ra`y(f4TvSQIhE V R R R PH G F A q UvX1yFt0`Xe41c1$50Ryex$"`Xf`yeXc1Fy%`Fz"R e1XR``XyR(a`"Ro`"UXRU1UR1"e"R Xpy xxe" cIyRexcX"eyex1FzFTSEe1Xy4Fyy`Rc`1"XT 1 q xR 7 f  q ~ U c 7 "Rl10RyFR""extcIyR1" IlX v`SE34R "@y1xxe"}"`0FV $ lad|1y1"s V A A q U@`1yXUUdyexep0j(e

Xr2)r%%5ge#aear#xgT)g5 #rXeeaXre2Xf2 r'eg5G#2 21we V 1fX#gert#XV`w V2 rgXF Y W #ref# #gCFewreyg2XrG1ygXefA# fdg#xgegre2gAXbge1 V reGr2#2 V

52 #1T8t1e#2ArX#2AdreX2Xd2AC V#AT2#wegg1)g #t#e2eg52w2tr2wa2%2eea#AXXgg 5Gr22XgAreFgd2Xgx2er#f1egyg5 #XeG'g12gerrr##F5g5)#r#2e aF#xX#Geg2#%G21gye2rXrXaf#1r5fr 2 r e ge  A2 g #r# % G Xe 2 # G e G agqeg12e2qaXgg#t1#r#22#d#% g15Grr#reXArg1gaXrerFre 222gbybfFdG #Xger2dF2X2#X22Xar2 rX#2FG22gAX521AX5ge2#2ryT2X##g1f rX g 2 a # #r A2Xg2er#CgaXr2#'eX#geg2 22X yge2g2A5 #r%rFrT2)#2r~req2%5) 2 12r2#2Xeg2##GXeGXgerreA f52 X ~1 5 gr 2  X g 22yb # a r ) )r GV a)AF)AgtV rCXe AX T1Gyeb%TXX )g1td2tf1#gg2 X8rea1Ce #%tAg 7 6 4 5#%8 Tg5g)aX2eA##21e2 Cf fr5f# x G 1 ) 5gGdF)G 2Awg2X2tFf2X22r #CVgX #Xx5gegaTXr#X#wg2X25are2 aFtr2#2 a2AC##52G#r22XX1#f%2XT f5~TVy V V g5)fAXe)1#gexGr#2XgA2er%r grX#g1rXTXG22gX#2CrXt#2CG221X#2 #g  2XGeex1gg5g12F)X2# Gge 8geGar2fFGgAX 2rG1arG C%#2 g22q#gG21F2V#2#%#2 Ce2egr# eFC#2w1yreTgy5~ry#rwyr#X#2 221#g221F#'Fr2AxG'2Xf2eeGX # #2d#2F%#2Xeg%2erg2XfXe 12ag#y#Tr22tT#ag1eg#X2%Ggee rqGr2Xgq2ryf25e#dCFewy1GXCF)%re 3 1 0 (  a2T#8~)y' V #yrXte21d2 15ge1f2X1G2Xer 2A12rr5frCFCre2e 2XgX2eeGXFg1#XT ef# rXt#1efe#X2'X2X2#Gegf2e 12gAX1XIgrX2 wr2raX# ef#)d1XT2erXe2egr#XXrXexr T1Xag#X1e2tg#x21#x#g1G#eeg##2 2IrV1ge#XggrCerr1e1r#25rw AGgr2X fr2rF5rwGgAXCF ee Cr#gr5frrXtre1rXgAw #AGrgbt '#e1XXr#gg#A2%aa 2e#2x1efXG1GrX2ergr2gbGbfyyr ~aFaGA r%e1TGArX#e2d212dG2ee '2 g1)g2Xf5rXXgre212#2%gre X2Xe51er#AGrX122A5X#%GG12AGXg X1e#2reg2Xf5rXXgr%tg2212wtT egeFt#%#ee C#2#%2 #2XrX2# 1Xe1ayG2yX1f5rXXg%eCef#CFewrerX F)2eg#A2rea1Xrt2%#2eGT2# 1a#1egC5r2A#5X#%X##2GT52#G#r 28a #~aV V 2#t2#1##g#212XrX2XrXgr#g X#ge'521egeTgA1agAaGgAX1)  e #X%# #  1 F #1X#rga#rea2FfVyA% 2gCX ye2#e%#r#2eX feef#r2gX gAC#dX TX rGg1f#X e2 eye52 #&xX% $22g2Xe511e2eaX21XgT e 1 ) t r g 1 ! ! 2 822gAX#2#b#dG" G1Xrg" #G r2#G1 " " ! rFXerTT#F2TwA#C12gAX'C#g #~2rG2TagVr q2%5X#%G22gAXFX#~2 1rXg5a2 i f k f B e 9 z c f B $6 e $ p 96 d e } 1h 1l g fGm gqI%qwGIF~| { 2%XgG1e AX2A'GgAX#2#bregrXXrAg 185i 5@ g11h h k iuj B f h f 9 9 e $6 2%T%AyIFd hhh 11y #yg#g1FTGeaX2eXAXgTgeGge##rXaf# kii f f 1@ B gh h $ 76 v p e z x V1%8%%%yt v @ fh fh d c4 "3 0 q p op0n( 1h 2 g1@ wrTP d5F1) GFw gTTg1F p %v rg'tre#g re'X12g rXt#d2 k iuj j @ f f 9 9 e $6 2%T%AyIFd di@ 15 2#2r2Xrw2%rXXrAg###XregTXrg5)12g 185i g g1i @ k iuj i@ f j f 9 9 e $6 2%T%AyIFd u t e w22AX##5efAX2X'rG2ae2#V a 185i g g1k @ mi fl f@ d c4 "3 0 q p op0n( 1h 1h g1m srTP d5F1) GFw gTTg1F g2XrrXeG2#%G##XrAg2r#g#a#rX k iuj f f 9 9 e $6 2%T%AyIFd m@ 15 f2XTa2#w#212e2regfeg#g 185i 1 g1m @ k iuj j i f l f 9 9 e $6 2%T%AyIFd l@ 1fe ag##TT%T##X#geGgAX#2#X2X2t#e%tXr k1i8u5ji @ g1f f@ f 9 9 e $6 2%T%AyIFd @@  ef2a%e wy2Xr2GT21aex2X#G#e 185i 1h g@ @ c  awg#r%F5Ge##T)8f'a%gT2#g 4 5fr %a a15I54 5tyf5dG))bG86 4 p (3& 4 D P4 c 0D& P 4 e) D V    g1#g522#'%re2#a1aer# 2eT2Fg221bf~1XgGg%1Xgxra2XeaX22 eaA%8T)ye1xc 1( 20 1)e) G1H GwG)V1( )D I& G0 v 4 h B@ 9 76 4 ED s ( &4 H P D (&4 u E E 4 ED s ( P u3 D  s( 4 r  P &4&& p  P i g&4 h g&D P4 c 0D& P ` Y U S R a21arA1XXr#gxF2X1rwg5)#t2ferX#X etGVqfdTq5IITGTTGe)e)D T1efedT)IbT86 aXWVTQ P D (&4 E E D B@ 9 76 43 0 (& $ "     1( )G0 1I5)H GFCA%8%5)21)'%# ! 
246

SCTRPHF)CBA9 USQ I GE D@ @

GA#GeaX2edre%GtGXTeArXC#21ee rAg#rGe#%yVXrG1#T1# G2X'1e#2e2f1#'1#a1 2#t1TICT)tredrXTX#rgaFXqg#rX12gAXr% V 2agVffq V V

ANNEXE C. ARTICLE THE NATURE OF IRS 13

i cR V p i w a cXi VR `i a c o 0w u( wF"o l tF11t(d FtvdbrBu$d"1wBd4hb"wa sV g Q QR V y c oa c Q y` c Va iac y ViaR ` Q oa ecR t1WVwawF}"wWgt"rw&dw1dw1SR"w0w"c ` p i p Q g V eca gcRci y Q RcXi V "Wqs10dF0dnWR0d"1R c oa s V g Qa `X Qa gc Q c o pj w gf e gca gc y V s gc ` "wt1WVwb}{ dFYSez$4dv1d(hbddF1wiuFtda ec e e` T i yV s acR RV c oa o Q Va Va c Qa 0Y"1xSRQ U 1q 1diwS"w&4WX"o l dv u( wu1W~wb` dh1W~1dtu}t@SRdhWgdSX"o l 1110(nhbxdwv| Tci gc Q q ci ` ic y ` y Q oa Q o Q j w gf e aa ec o T | i e g Vi` Ri `aR eca T V Q sV R g Q QR V y c o 0wSRQ""y 0w  Fwtud1d0dbS`1SRktu1WVwaF}"wa y V s eca` QaRc ci` RacR RV i T |` c o p ecaXcac e aV g Q 1diUdb@Wyw0$dtv4nWc"ts"vddYEt"SRBq eg` jww X tF0FuY i t` i 1V@F~1kSR$d"1wnWc1"gwbwa eg i s gc Qq Q cXi VR Tq QR ` a` o a ycX c j i gf e aa `a ` e x c oa s V q QRR cX Vi y c "Y4tF1t7tdwh4bh uuY dtWgd0Yw"wi c oa | ecaXcac e ci` Tc { T` yR c oa sV RcXi VR c oa pj i "wBYx0w0wYwtxSe"vWTt@$dt0d"1w$"wFWTT 4v110w Sec zvtuddEtdWRzBWgvt4Bvt F d0wbS`T Q X y sV c oa Va cX g`a Q e ` q Q `a | W Va eca R g `ia p i T | ` Q gc Q q ci ` Ra gc g V y y V i Q ` (tddY cWtWg1W~1xw1z4FF@1X 0w u( Wgt@y c oa gc ac R g Qa`i` ycR i T q g` ecaXcVi y c oa R g Q QR V "wc l |1WVwb4twUtS`Ftf0d40tw""w1WVwavy cRc oa y Vi p ` y p w Va ` c Q i Viic g Q QR V y Vw c Qa Tc w0d1dRF@ 1d}T1"F0WR1wd1WVda wF}U1W~wbS`wi c"w@04F0cwFFqdp"ow1k0d"1wp"wBtuw0"wF1Wqd| oa RcXi VR aR a o Qi | c a i V pRcXi VR c oa sV RRc ga o Qi c o g V R e gc yc e q ` Q e T V ~ g VX e c oa Q R g Q QR V y c oa s "wa1n}Yc11@Wy@0c1W~11vF0cdWg1WVwav"wmtV g Q Xci y c o p c q Q e g` y Q yc c oa sV g Qai V y c o 1WVSRQ0w"x$11t`@Wy"t|hYB"WgQzdt1VdFx"wa sV g Q TV~ g VXc e c a sV R T ci c o yV ec~ Qiaci R i t1WVdaYW11vF0odst4a "R0dda1dis01WcwwdE1` l dwa c y ViaR c o p i T | Q gc q ci @1wd(1`@$UWct`spWgF~QFwt` TU i1q 1widwSWgQ yV s RacR RV Se"0w " Y YduWgv0cd"1wuc"wastv1WVwawFz" Tc { i  p Q p c oa Q R Xi VR o V R g Q QR V y c o G E C (1H(tFDB "
I P

p l WV}0Wgt"SRFWTc l Tc | ec Q y`c Q R` T R ` icaR TX c oa Va q g Tc | Va yccR gc oa e g ` iw e g ` i 1$d(Wz"wd1WVxd"wUt d"1B 0w iw | c e` y Tq g Qia c oa Va cR TX ic~ ci` c o p i 1w 0&v1@Wc11W`dwx"wuwwFWVw1dt""q 0w

e g` iw e T`X ci` g Qa T V~ gVXc e | e T`c ~ci RcXi V "th 00WcT1rdt1WVwF1v10Yv0c1Fd41wR ic oaV p iw e T `X Q iw s V oai V g cXi VR a o Qi dtV l " 00WcT1SRri 01d1"p4FF1Wqw"| aRcR TX c oa Q Qa g VX (0vVBdt1a "vWgwv1@ p i 0t` i 0dv"1}1X iw eg iw Ra gc gV y yV V l dEWTtc l v$s1WVw1Fv10Y@dabWc"1Udt0}y a c oa T`X R o p g Qa TV ~ gVXc e ic s` T | V e Ri`c y ` i w e ca V g e ` o c g V c o pj w gf e ei ` y ` ti 00dt&F"uc l "1"4117khbBdr4t@"t Wgh"1@11F0(7tUdwWg0{ dFYSe}fYFdWit1o Q e g ` j w gf e aa Q e c Qa g c Q g c c | e ` c T ` e ` Y 0u"1kw 0141w(0wv1Wqw"nV l k"t41t"Wqs} t iw e g ` iw RcXi VR aRca o Qi | a c o pj p Q p Wgd Qa l 0Wc4Wi"440Y110Wc@bt@w"Wu11&` Q o e TX QX gc RaXc | V aR ` T a` sV icaR TX aX ` y y VX Va Q ec T VRci Q cXi V i c oa g QR TX g VX wFWg01W~10d&WRh4FR 0w u( dz1WVw"W"1g  p Q Ti T Qai ` y cXi VR a o Qi vWguWdtS`SXw1"Y14FF1Wqw"| ic oaV g` sV Q Q Vi y c oa gV q Q e gc yc e Tc { c oa Q yi Q g w1t}tB(aWyYddwx1zWgYYWek"wWgw1Vs" a V g Q Q T Q o p T X Q e aR V y c oa Q g QaXcac e c oa cic tSRnaWyWQESR$"au$(vBdkSRFVddYBduwo l j e g` y r p w Ric T { c ycia c c oa i s R g Q QR V y c y VR a s&"tWw1dwawdvwFV"1WVwawFu1wUb` gc Q q Q Q T ic y y pic T { X QXc yR ` Q T | `aXcac e a V g 1~FSRa @yWQ}"Ug da v}{ 0}zUWgWc1d0wY1sSRQ cXi VR c oa a` oa ca` Q e Q T |` Q R g Q QR V y a y p e g` d"1w"wkt"wndbWX"WgBuc"tsWg1WVwavUYy "t| c oa Q j w gf e Q ` q ` c e V c oa o e g p"wWg@t110hbUdWgtF17YYX pdwaQ l t` e T y y` c | aV g T VX g Qa`ic yV g Qa TV ~ gVXc e c oa cic WcWQ"t@}tSeF1WVdbd}11WVw1Fv10Y@"wUwo l Ric T { c oa i s o i s e c oXi ` cR ci RcXi V dda d$FV0FhBbF(pdaQ l 1V0"d4t0dc l 41wR cRc oa T` gc o p i p Q gV ec Qa gc Q ci` o Q Tc { T` y w"wWTt"""Wqs1{wFWeBdtdSXo l Se"WTt@wR SRdWgQ0w0ddtBt4twgu04FFc l E"Wqs1V Q oa ecaXcac e ci` T`aVa Q RcXi VR a g p p Q g g l "SRQ1t@Wy01W~1Fv10Y"thzWgwa"0w&" V oR c q ` Q e c T V ~ g VX c e e g ` | y q Q T R ci c o p ecR Q y c | g`X e g` 0wwSR}t"t| c gV Q gV T |`aXca e RcXi V s ` e g` o ic oaV c oa "1&WgWT"1@Wct40wcY0d"1R l cmt""wtdg  pd1`swwtug1WVwW11v10cuv1VWwbg@tdv1uSe11dd@dFWgQ RaX Qai` Qa TV~ gVX e | g Qa` Q y`a gVX QV~` Va e gca c l t"wat}@Wg@0d40cdz|d04FzdBWgwWi"F0wU R e g` | o V | Q ecaX ac e c Va RcXi VR c oa q Q Q ci p ` ` e Q Q yc oa sV g Qa TV ~ gV c e i s` y e g 4ab}Wgzzcdt1WVwYW1Fv1X0Ykcdab@B@t` x guwt}| Q oaV eca cac e cR V oa i` eca T i `aR c T |` Q eca gcRci 0wX0wY@F"wxcdt0wWRWQuRd1d@o$pY@Wc"ts&g0wv0w"y SRu4cda"r"wWgQ1WVw0wYetdcdastaw0wut"$ Q Ri T { c oa g QaXcac i`aR o V T Rci T` g { c o 1b(11HS

& $0 b 0 '0 b 1 9 9 $P bf9 bt b 9 # b" $0 t   ec Q`a|V gcc | c~` 0"gt4YFz1do 0 " $# 18 1 $& b A # P b 4" b 0  4" t) # A3 a gc y g V Q ~ gc R Q e g` T y y VX i c oa g V R T Rci q F"1wWivYudatWcF0w  "w14a "0d&WgQ d tA '& b 1 b'A b $ b 40 'P t & "  1 1 9 # 0 99 l WVTFV$"wttWc"tW`WTtbd4bnSX1dFwd}w"wUtdWcts7 T s c oa T | Q`~` `a` e Q yVXR ViaXc yR c oa e g` jw T |` 1 $ bP t# tA 9 0 ' tA P 01 bh1 09  $# 9 9 b8 s3 1  P y r p i e g ` p w gc ac | R c q ` Q g Qa T VR ci o Q o c oa y Vi s "t$Ww1c l }m01t@Wy1WVwW10dhY1Wq"und1w" d& b 9 tA b1 b'A b 0 9 b $P 9 9 9 # 1 # 99 A tA 9 tA d0 bh$A b 0 b 01 $ t 9 1 P P 9 #  P 0 b$ " 'A tA A 9 A P P 0" tA 09 t0# b 0P b t d # 9 t 9 9  "  4" tA # $" tA " A P 99 t0# b 'P b 0 &  d d $ t 49 A P b" tA $" tA 9 th'# b d# b $& $# t 0 P A # 0 9 & 9 p " a ` i b i c 1 "o d l R T { wi o Q ` o V &  bP 1 'P tA " b P  @A b" t) 'm3 "gWQ1Fd@wWgt4Y1w0WVe wQQWc b"41z" ` 1WVWqwUQa "WgWctvt"Ry c T e ` V i | c o a Q ` a | V V a T y y ` F l g Q a g s q Q da V V w R X o y P 0P tA '# tA trb0 b 0# b $0 09 t 0 A A 9 A 0 9 9  99 pc | X T`iaXc yR c oa yV s ecaX `ia c R R y sV g Q t}"tdd}w$d1wi0d4Fddvs1` l dvtt1Vda # $" bP 9 9 A A & b0 tA 09 th0" b ' b P  A b" t) m3 T VRci T Q Q Q c oa a` c q g`i Q oa W10dvtS`waWgwt11t4SRdgQ 0w u( tdw0wR i sV y iaXc y 9 t0 tA '& tA P th' tA $ tA 0P dd# t  A A 0 0 P 0 g QR Q yc c oa y Qa g VX c oa sV g QaX `ia | R e g` q QRRcX Vi 1WVdWR@d0"YWgwF1dtxFVd1ddY"UtzWgd0Yw"y 0 $0 tA A A #  0& tA 0# th tA ' tA '1 d 09 0 A 1 `a` ei` e g`aR c oa ic p ecR gcc | V T` c ~` o c db"e uk$ 4ttdd$das w"UwSRtB1d""1WgWQT & $0 tA " & tA 1 A " A 1 $P t9 t4 tA 9 & bP d $A t # j y r p yX p q c oa g V e cica gcX j yX q A 0& tA  1" 01 th'A tA ' tA $# d '1 t " A   # 0 P sc x u0WwYhWw11v"wrFf0wwvx1FF ic T { e g` | Vii` g ` oq Vi oa XcRi` y T`ia gc # "  $# 1870 9 1" 9 & A 0 A 1 9 1" $# th t9 tA $# d 0& t 63 1tvda t"H l dwt1Fw"wutwdtpt4wFX 0 A " 1 $A bP 4" 1" 0 0 'A 1" 1 "  9 15'0 tA b1 tA 9 A 9 0 bt# 09 s3 "wt"10EF1tWg0"g1dYFdb"e uk$ dn c oa sV c g w Q ec Q`a | V `a` ic oa A &  bP 4" 1" $" 1" $ 15& td0 tA 4" b'1 09 3  1 " A A  A # 9 p g Qa Qia g VX i TcaRica Q c oa y V s i TcaR c o FVdY"|dwv1$tS`WTwdwvg"w1wi$tS`Td(da & " # '0 12$A 1" 1 ' 1" # A  $& th tA ' tA 0 0 'P d 'P t) #   g` g Qa` Q Vi y y` c y VR o ca`i ` ycR Va T QRR V y 1d` l YtFVdb@WyYdd"tB1wd Qa l tdbd1"uwuWcW|dF}$aQ ( 00 n '$& s% $# m!0bs!0 0s 5 1@Bvi FY71WVwW10dr1WqwpW"1ctw"tS`T  9   9 P " 9 9 9  c e ` y j R y p w g Qa T VR ci o Q o c oa T g p T { Vi y i w$"wmF$w"wvgsFSRx1WVdWR@t  dEgwWc"1w"@i TcaR c oa R ` cR gca Q R ` Q g QR Q yc c oa cX QR T { Vi y % 3 ) 3  3 323 %3 c oa i s ca ` Tc e Ti T Qai ` y R g Qa `ic y V Q o p T { Vi y i 4(C vwFFn(bU414d(Fv4F) s "w1VwbSXWQYWwtS`"SXdt1` l 1Vdb4}1SR"1tWctw"tS`T h(Fp(h % % 2 Fvm4&$71d41mvYnUFpI H4FI wu"w10wF@yd"s4jtt4wFqs7udv"g C% 3C D D A9 8 65 C 32 3 I 3 DC %%3 T caR c oa g V e cR V y Qi c y R T c g ` y T `ia g c X p Q Ra g c V y y V X aQ X VT c ~ gQ ` y l a s V c i c o c e ` y g VQ a |Q i a g V X C D G % 32 3 ' 3 ' 3C 3 % 3 2 ) '% 3 D % 3 b0G 1FI H(w41d1wbu(h % u($d4(10(Fwu(d(C 1F(1@1@V }tw"n1@11WdY"dwv1  c oa c~V yci Va ecaXciiVX c | Va e` o R T {Vi y c T c oa cX g "wn1bww0d40dw1B}wu1"E0ctdzWgWQBdF1"c x v$@vHEv(vS"0th}UbvYd(d(C vnEFvh 8 #5 3 DC ' I % 3 3 D C I 3 DC %3 G 3 ! 3 DC ' 3 A9 8 65 3 D pj w gf e ei ` y ` c T y T c o y r q p c oa Q R ` T FI td4 rkFI 4EF(0(Wz 4vn7 P b P b$F(C 1thbkdd1@tvuWgWQt"WQ"ts1 YUdgn1mWTc l R ` j ` i gf e ei ` y ` ei Q ` e g ` Qii V ' 8 # I C ' I '23 3C23C3 %C% 3 DC ' G Fu0' (1) 0dt0d41tu(w4(d1u(hzv(s}vF' I 1ktF11thbd4t@"tv11td1W`WTWTttESRwd1ri P txv9 0ud47dYs(b4d1H2 (Hd0(FD "uWF Wgw"wvgFVwwSRxtS`WTwdwvgutr1WV1dphgfdbYWVUSRQ A I 9 I33 C3 %C I3 3 )% 3 C3 'C ' Q cR gca Q g QR Q yc i TcaRica Q sV g Q qci ` Q eca`X T P I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$" 

247

2GjUYGty2UiAtW2fY F2 w p2 c2 2 2 15Y2s2#C#t#fi aYa U1GfYfCiia Y Yf'UCfi5)Yciw)h#2C15Y2 G t'2A #i2 j#2U2jW2fY#iY#f#i#22Uf51fY  D 9 6 1EDeAB@7c4'IF2dCYhF1 5h )52GQ #a i22i 8icpwth1Y#t2vfY1#2# h#vipaa wUYi U1h1# fi%5 #1fsY UiaitGYh##w#j2Af1f fAai1#2%SfU#if2#2|U'A2aai GcU5U#% t#i f aACY Yh U1#fG 15YGCYf#UiiF1Y#UF2U2CYfaic##iayaiY Uf22Y#fi##2f)C#%2U1#vai U#i )2#apaa

 1  E  i T D  i @ 2 1 D E 1 D E  i  HG 1   D p  D 1  E  i D  G D  i E  i FE   E D  i E  i pE @ D  i E   D i     E  D    2   1 D E E i  I E HG  1 E 2 1 D   1 E @ i  & E HG E 1 1 2 D D  @     D   D F  D 1 E  i  E 1  E i  VG  1 E @ i  G 1 E pD i  CG   1 @ 2 D   ` ` BF` %5d  ` x y y x E v & " q 2 iw2 dA1SG)Sf5)2)%GWru2ts)fIYrpY

hf d b igeca

` E i0H))Sf)5fG)pt5)2I ` D Y X   T U WG)@1 ) )1Vh shh)%)5If2U 05pSsvI1 T R  P IE GYG s)f2w 1 )ssf)2f)5fG hQGi G1HG &E G  EDD 97 64 2  ) ( & $ " hHhFCBA@85IF310G''%#1)U!52     ss1 sf)Gw1 )5) 2SGh hW22 1 5 fp t2 GG)F)5I5)2)c)G%1s) S5 5II )ffp%G#1 5h )52ScG)) ps 5hv i d%hY#1aUS

pW%#j%2 F2 v) 2fGWi#U2#Y##22 # w2paGfYtU YUfyF|2Uha15Y2Fi#2A1UGUi5 YYAUfGf)2jYfi#UYUYtG'c 2UaU1U Y #i2#f2G2UfU Y Ci5YGp5521vi 1aYfGht#G12af2f)s# iF2 v)i1if )1h#2t1Fhf2 22 # U2#a 2 'tSivG #5GS2 A2Uhaw#fYf# h 8ipji#iGi%U 12af5Y2iWF21f#UYf2Yf2f Gtf2t f#2p5tYfjfC51W1f%5Y2#p2sd2AYf #U#21  21GU#s2c'2Uf GW# i22i p2YiG2chiafGYG2f2#t2y1ha2fYv e q l u ed b e bdm } e b sm y e Yb Gs15cGdpf5)21IFf1x 2C8 b r g }emv r gv y b s r gv r gv y s s rbdbb r s qv G)jhh1)itG)ig ihhh G)ig i)ff2C e h q5d ))ss21IG# G5)21I|1y%8hg 1x th)v)5I5)m v smb u u v b } s r ~ e bdm } e { zf xw q v s sm e b gbb u lbm mbb u s r r e o e qd sd b n emb q g l k b s qb g f e s Gf)tf)sv2pr q iptb fm 5wG) hj%wh)jg ir 5hwix d b e b bb b r ed b y bxv s rb qeh gedb ` V XV R G)fic5Iw2i5Gt%1wu t5ph 1pi1f5caY1WUTSQ

P D (&4 E E D B@ 9 76 43 0 (& $ "     1( )G0 1I5)H GFCA%8%5)21)'%# ! 


ANNEXE C. ARTICLE THE NATURE OF IRS 13

248

A rq P P 09 P v rA P e 0$  P 09 v A rA 9 9 jpA e 09 v rA 9 P 4j u09 v 9 rA % h q p% 3 h ph r v A   CC h' %32 )' 00TrHEwgj4e(F0H% r A9 8 65 h 3 DC g ' % I C ' h 3 f ' f g ' 3 ) f h ` U S R 0r7oI pF(iF0' iel0(Fux1l( C FEt XWVTQ "F "F@ yiFTae n0e10eqe"iFx1e 010c1eyr$"F0n"1ir10F04FuyT1 "qecq yBTnype"1o0Enj1e"jxjj"4la1e T1ec1 Fajy41q" y10@ " Be y"0a"1"1j(01y1Xq"yi 1 xcFy"qu1&1o(T 4 y""BB$"B" 0yFy0y"xe1( 1o c 0alin"4( }pq1n01pa&1 11" 041y @"1oyjqc1 1cFy"o"yjl"e " $E0lcy "1n"F e gTy y11c e4(ly0"reFX$ $01( lX1 qe 0(0s TXiy1u10c$@"1u1eFly1"je1

#A0W9gTi"9"W%"4Wse@Ws#d x ~ k e s k x f s |d t x d e t u r x s k 9Sg4 `#6f WqWs""@eI9WSh%i9#d sidI#E9"@W9e#W#"%jsi9"sWh"EE d r x x e r x x r n x x d x ~ d s TdSbF#WbI"9#%E"x9B""W"8i fd | ef k r s | d t x `Ag9e9jF4W`W#q#4 `#6f WFW x x r i u se@94}iT9A92HsS8T9IW9wF694BA x s k r r ~ s |d t x 9ej%FW4EE%8}98i%#6if@W ~ x ~ r e e n us |d t x ~ 9`W99I#W4 `f WqWH9b#%d d"E#"B9#9h"WgAAw9z r x s| d d e s k r ~ x s k d 8%8#eiA#}"Y@WSE"%@#"sBW f ~ f k k k x x e e e e x sYqi6WWEWse@4A"#e9e#sW06i u9sy@`siA%2Ts@Fz"SAwY"@i@##4 s | m m t i ~ x s"9WEhe9b"A%h"9sh`9IrWhh##4 e x r n m e x 9&W"@s"TY#%WE"h`@#"9I}q"e" d d d ~ e x k x s k r d Hi%#6if@hWFiqAAWWS9g"e" x w u s |d t x d x ~ d e ~ x ~ x x 9d9e0&9#"#4zb9W##W"Bs l d e r d m i x w u d d 4W4@"s#`IiWEp&W##""W"4 f x ~ e r s x ~ r r x r `gWy9W9g0i4e"AIWw99sIIg"Y~ x e ~ d T W4"&ibe%S9#4"#%iqWhs"Ww9Ire" f d u u k | x d s d `EqeWHee "A#`#}ezWe44 `#| W`i|W9#eFT%&"WwAIBHW9"W i u o n m x ~ r e x ~ e r d r s k r 9z4s9W9&A#WeI6f "iqB%ys9Q29T9IW9w s k r  n y x r ~ x ~ r { y x %S`AI}"sW&bW"YHAS"e9I4"I4ybW ~ r x r ~ d 9`s9e9yd"sI#%99"`S"Y"TT`y9Ir ``%WHI#zASA&Q9EiWid%}WI&hzy x w u u m i ~ s | x { x d x e x w u e t x WgiEe9WW#qs"4xT9eeW9wAbWy%v"6if@bW sIEdqB4#FopT`Td4%AF#9#j#Sh"E4 r r n m l k d ii g x d f e x d x 49eQ9""A#dB"g""SSAW9"T#AWyw v u t r d f p f d c a V XV 9QIs`q e`ihgeHb`YWU 3!3 3 3 ! R R P TxSF H3 4#E4A#pQIr7 $ G  D  5 5 ( 7 )! $ (3 ) 5 %#oH#FrxE"C B FA9 (4@94T8 607 F# F63 $ 1 ) ( ! $ ! $ ! F4%#&0p g3 F"42y0e%"("' 1&4# (%#j4"(     " Tj 0ur7T0p $i@uHr4Xe acX

v 0 e v vP P 149 9 e v 0 09 v v 149 9 r v v v 0 149 P 9 e 9 v P v 0 A  149 9 9 r e ve e v ve v 0 v v 0 149 0 9 9 v v v 149 9 9 A v vA v 9 9 49 v 0 vA v A 9 9 A 9 v e vA v 9 9 9 v 49 v 0 A 9 9 0r vP P vA v dA v 0P v v 0e 9 0 9 9 P v 1 9 v  1 9 9 0r vP P v e v tA A v 0 A v v 9 0e 49 e A v e v9 9 v 9 1 49 P e v A v9 9 v 9 P 49 v v9 v 0 009 A 49 v v9 P v 0P 009 A 49 r v 0A A v P P 0 009 9 r A v A v9 vA v 0A v P 09 0 9 P 9 ur ve e v v 9 e v 0A v 0 A 0A 009 0 49 ve v9 v ~ { o}z o|y { z z y w 75 s v h ' g %C% 3 t p s3 I 3 D x3 Faw0(Xuxq4Tp(FrT(C I 3C% 2 r A9 8 65 3 DC g ' pC3 h 'C ' D f m I k i h ` f U S R Tre(xF) 4&04r7jF(qH(y0(FTonT lja"gXWVTQ

0r Pe AP v dA r tA "u10j$"exXF lx1y1qgTiyF1 0 9 v tA r 0pr E10j(p( ceF " 0ee1eF "yqe e0l1r" 0A P 00 P v r pr 1 e Xc11 q""e4esE(o0"ruu1u@c e 00A A v r 9 Br l14 iqy"T0Eoe1y"T F111Eo"y 9 9 41 9 v r 9 Br 1 14yyge @y"F0csE1x(E4F 32 IC v f3% v G I $teFyx% $w "4HT1Fut e4(1018 %32 )' a110}x(&eqr("1$"$" Fiyx %32 ) '% r A9 8 65 h 3 DC g ' I C f3% 3C23d' b ` Y U S R c sl4 gjj1 (qcjcX" e4(10Hs0$7qI piF(V@0' (T4H(eeH ca@XWVTQ I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$"  P
249

yqd11qi1uur 1Fi#VT ivei#uip TpT ae p T r e Q e Te U R p Ti e Q s yk r r e iT a eaT ip s p R UT g o g 21veh1x#iVa ViVVG22#d2uR}vfuVFii p eaT Y Rp R y ip Qi Uf T i f i R Ya eaTp f R UT 25C#xq`2qie e#o2uv1qqA2i2V#i51 q vf#AR { zy W xw g raTp f Rf R x b { n U fa eTf T i U g i R 21h)`2i #V51 q }2ivf'hoFu%5iViuui}2iT ep Rff p UT R a iTe R R a rTea T e r ep exV2veC#F2Vepu##e5 q Vj}xiCuRd U p e UT ff r Tea f R R U R U Q g i a n i #o2 pAixV1}qjVe5ie q pfvT 5vpuSx2o2a g R rY RT R r Ti ep y UT R U Q s r r 1uUvfu2#)vfujxdi#uijuR f #CuS#t1qxu`#%vfR s e UT R R R UT a g f T R Tf i UT R R U xie#ixr #j}2ivfuV2i px`mG#vf#Rvf#o y n s y iaa e e s y r r e  g 21`8i1#roe2`G1GetV2hes g G12e%ui5a k { z W x a g y k 21ytd)wF2i  tuR yqmS 1F}iA2iVexT e Q e gi RTep R R T T T e Ta a Rf e e ipf iaTp f f r R ya quh2yRieVGV}~12i q Vvvfu2#idvpR q VGV q #2iqVv # Ivqqx)'#u%u`tov%wxmd#v#

f r R UT gi T T RT f e U R Qs p e U ep i e #tqF##imGqGiveee q hVT S1xi##CiqidvfimvfR { zy W xw Y R r iaa rf Rf R r r RT ip UT T e UT R 21h)uqxu2V#XGivfu#``2#jexw1pVT 2ud|itI#FxYm2|Seut`u2i q 2xuG2Y 2xr i r ep Tw R l i Q i Uf T ip r ia p e vfuUVo121iVuVpi2i25)xuCTeh1qVa q 2iur22y R T f ye a eTf UT g i RTp eaTf r rf a g e ffp i RT i p U ff e r a UT RfT ri g fe Y R u2i1qtu%2e q xi#VT#uqvVu2vf2e}2GX1#2iR U i g a n Y iaTf UT p Rf Rf R r r RT i #T2Va uRVou2Vwux2ivf#u`VG2p iT r R Rf R R U f r p r T U Ra r UT r ia #qX#2ivf2iVvfvfu%p#x``2Y#uV'xu2Vie 2iVfA#rtRqSe##ix#22Addu% Rf R f e f Qi Uf e r UT R Y e R u b U xq#wx)'dtexqueAi##dix i1))uP `15IVAP I15t2 H 9 ( D4& ( 34&3 &(3  P  4 &4D ( (3 4 E H 9 ( P &( 4 ED & &&( `SGF)v5)`P 2e5GpiFi1( D )uGi)jI1)V3 4 E B@ 9 76 ( P D 3 4 ED (D &  P ( &4 (3 4 ED G2p15j%8#j1( )5e( vGo2`1q)V))VC8D i53 GP Gu( 4 H 4 P4D&4 4 ED D & H 4 P 4 ED D( & D&4D E 4 E 12S)IV`)x%5 G1( G)SP tuItVe1H 2mGe8D 253 GP F`)#)2)VG))`)dGP 5`IiV)v5I&G P j1( II& }VI 4 ED ( 4 0D3 0D& 4 ED H ( E& D 4D 44 4 P & 4 4 ED D&4 H 9 ( P D( ( 4H 4D3 0D& P(34 G)P vIV21( )IuS1} )VG)IIG1V57 u1v
6 5 0 5 10

4  (3 9 $ 4 ED E P D4 ( P D ( P(34  4 ED 4 12iVpAdG)D v( )Vq1A( )G0 1i15V2SG)FD  P 4 ED (3 B@ 9 76 4 ED ( P & 4 4& 0 ``~G)}P 4 G}1V1jo8G)xj( I)& }5vIG# G#1xv

e a e r e k R y ep R R UT R U r f e2z 2 #eFt5iAu2p)T`Iujutx#xoqve U R U Q sae n UT Rf R a i iT f p T r r s R yep R R UT g Vvp#%1i%~uVAu#}2Giq#2)T`Iu#oii Y UT Rf R T ep r T i r i Uf ip { zy W xw U 2u1%#oqvxVeqi%vi#vuGu q 2w21td)SuVT Te UT a g R fpa r if { zy W xw UT i g r Y R R R U e#Vo uCwq1V#221ddCu2a qxuu2iCuVT Rvf2ie5T#diV#2u%m1iouV}e##i#ie R RTe Rp gi pa r if U ae n UT s Qi a UTi gi feY Rp ep sfeY UT e R yep R R UT gi Y iF2GVT#iui`GGCuV#2A)T`Iu}uVidGq1 UT sT U Rf g i UT i f RT r RaT ip i QT g i R R e U #q`2Ypi}#R}#2ox2iuurVG2Ao)mi2iTq#2juVT iT r R { zy W xw gi R Rf UT r iae f e Y R R UT g o#qvf2dxi2iTGi q ###2iGGj1#2i%#xii ff T U Ra r yaf r i UTsa ga U p e U ep fae 1#G2YVuC12V#GFuV2GiVu%t1#2##CiqiVi1 q q 2iT`i q 21h)j#t2tV`2i2iAA#i%n e R Rf { zy W xw UT i aT p R Ya e f e sa e UT g e Rf T R UT u R yep R R UT gi Y UT T ##ii q AjT #VGo#jh22)TqI#SuV#io2q1%uVee p i Uf fe Y UT g R k pae T U Ra r UT iT Y r iT f 1 uu2GF#uii g #YmlhVim`2YVuouVSVdu2ixhVAf { zy W xw Rf e r e R Y i a i e u a e n UT iT f z 21h)pvf2#qipG2iGuU q 2A}Ah`1iohuV}2s 2I T e sT ip Tf Q f UT s R yep R R UT r ra rTa e1Gu2i q 2C)1oeiA#2)T`IuA#`d1# q f e Y UT g i RT i UT iT r R f T e T i 2GuVi2iViuVjhuuFvfS mig )oeo`#2i q 2p Tf Tf g U eaT pk R l i Q i Uf f e Y R )12eA#op5u2e q i5`8wv#Ym2XSeuj2G'1#2iR UT g i i UT R fT ip R e i QT UT p r i ##iF#2a q uRo#5Gu2i q Gi~do)F#F#uqa q fa rTp raTf Rp y i QT f UT { zy W xw g i T U Rf g i U 1##Vt)Tqi2~o)}#|21wh)vihG2YviuRuVT u a b a UTa i s UT gi Y T U Ra r UT iT a r R a hGtVd#GSSuV#i%2up`2Y#o#pie#vpq# q q u f T1`R1Vf#Vfujx%'2ivfSf#it2Y2i #2uGpx`V12ohuVj1i2s g #Ym8S)1Fq2uj5i%eT 2`f p W f rp R r R Q R Rf R R UT g i R Ra U Y r ia f r UT s a e n UT e p R lk Tf Q UTa i a e Q i h g U y r e Rap Rf R UT r p eaT Rf R T e UpaTf s RT R R a rTp raTf Rp y a UT i b R #oi22eix21xrVVf1q2ivfGep%#}qj25Givf ed1u51xGuCvfVuV5xV))Tqi2Cu2ud2iVT f e Y R T U a r T g i e UT T ea ip R Rf R i Tf QTf e e R e y e U U R U Q sf rp g f e c ba Y W U R 2GA1#2iRvG2YR#duUtiFTe#V}io#ri5ie q G~AGivf 2)1F)21#2A2xwVvpuSt#qi h2Ad52`XVT SQ P D (&4 E E D B@ 9 76 43 0 (& $ "     1( )G0 1I5)H GFCA%8%5)21)'%# ! 
250

2 0 2

Tp W R ff r q f 1`1V#p R R U R U Q sfpa r if T a R i QT iT Y r r i Uf f )vfudvfVvp#%#15u2h`Vu~FA~u2itv#2ud1#R i QT U a e r e k a rf e e e f e s F)h#ou5y 2 q8ip5iAu2pVu21u21aio2q2#R U UT a f U R Q a ip f y urRu#C2i g m2 dXVT%d12e q 2s mg q g l e UR Q iihm'|TSt#R qhmd#i#Vt#Roea UT e UT a Qi a Up r e R a n s r a e x R { zy e { z iex~5#ovfFq2#ie 2A%uu AvfR f 21#2}F21y fa Tf f R i QT UT gi RT r RaT ip RT a UT eTe 5ieVuRC2iRVfAdFpuV`iC2iVq#rV`2v2yeveu%5eu x b n UT g RT ifa RT f UT U R T f { zy W xw U h{%uV`iiC2iVqr2Viiveee q FuV}T|1ii q 21h)v#T gi RT p T Te r rp eTe x b { n Upe i g z i}2ieeqi#Ueesi#dViuXh%jV212Vat5vfuR q zk Rf a rT e e UT i TpeaT f i g dVuVa q dAifuV2AV2VV`v1#iVa q #R CiG q duVxiFnievV)uV1`1a q vft#Ymhu UT e a a Tf UT T f R R l u)j#%e#Gxq op#v#

ANNEXE C. ARTICLE THE NATURE OF IRS 13

{FW1{"bnU@110oU0oo`on4( "o1"Uw"U"{Uw1@F0o(b01bn4"1n "{fnFohb{{11hWn@oF0os4{o{Unt1o hYE{F `{FE"1Uwon"F{d4WBY 0{UU""U{F0{(F"{}1`{F1ohUU`{1E0`1  r q p s  ~ | HBYpo}TywuBo&Xj { z x r v n t s r q p n m l k 1FoU"no1Uo40o{1}"{01n@o"$bn110h Y"UE "n{1hU"1""U`"0{ 4n1F01U$ 0"o(Wn4F{nE0`1"{WE1"oFB{1nBo nFooU`E"{UY1U{bow" 11"noU"{1h0U" nW{1"U"0o`f4no@0"{S1b`(n4{0b{ "{@"U`b"nr1F1mn"boB1U41 UbU""1{o{ y s c E s e Uyi 0C F w1w&s "{Uoo`4n4(doUnrn{Un"n`s"n( E x v i Whow"1W1d0("{bnE`o"Bo{"W 1YbdU"n U0{w{0{"sow@{{10$Uo{F{n" h D v v v s HHHH h i q ggr pr p mh e "v h C g v 5ut r r f pg v v v HHHt r x q y"r pr ` `(1b(f4b``(bf (n b&hbhm`H1hfr`b e "v C g v 5u r r r ppph v v H5u x q y5i pr xg 0mh d ee 5v C g v Ret r r r i pppi v v H5u r q p pr lg 0mh c (H v 5u"  v v v s HHHH r q w pr G Bh 1Y(1o C v 5u r r r r ppp v v v s HHHH r h x q pr e h "U{hbU1o1@U1UUE1UohU141U{nU"1UE`4n 1 v 5u r r r h ppw v v v s HHHH r r q pp pr sh 0{Uwo{1 wU1U"$m 00{Uwo 1 HHu v 5ut r r x pyRw v v t H5us r q i pr ph C 1U{`b1@1U{nU"1U"{sWnb{{hUE{@U01U{wS 8 d g2 mec f Vd W b` 2 ` S Q P aTRYI W XU 2 U S Q P VTRBI y s hf 1Y(n1U{FWdoh$Fbn@U1h{n@o$w1(wY g u r k vc ys r r x G p "Fx 4p s$c ( u o (FH BHg 4c f "@onF`0"{Bh1o{U{1U"Uo1"`{n4{F u }4(HF4H FS4(10( k f F y oHc k D4 7 ws "{U1UoUoo`m 00W 0U1UmEhU vc xp v s cpy x psc i c E s pc w Cg f e v u cpy x h g f e k i A k i y c rsys q @ V 6 4 3 4o(F0H0 7 c  Bh f c ( f(oc {wF u ( fX 987502 "{n1F01hrY}whU1{w o bFnY "{U4h"B{ on1{ 0(BnFoFWohw@bn""@n r 1 " & h ( ` h @@ n 1 d n w {0 0 n41{410n}oUn4{"1U" {fn{w{{0 oU on"o{{ E1Y`Y Hs"m01noH4hUwn 0`w{h"`Y"` U`U oE&4hojbn 1Y1 BEo" w"n U ` 0 r ` $ 4 1n { &o n do{{h 1 n 0{w{0{"onn4(j41014{n4nb1 owFb hwmY` o`0}UU FY {({ "Un"{b o 0 W}b0o"`"BBo 110 S&{ow~"Unoo(b1@ 1 U"$ `"n E(W4FU1Wono@`0Uo{`UUn@$U ! {"{ 1 UU 10U`}o nd{"{h b 1 U "F `w 1 " o " {En 110 hoW{1U" 4 1 (' &%& 0 ! (' &%" h}Yd"1{10d1o{0b{m1 `F 1 ) $ $ # 1   $ $ # !   01{"`WF41EEn {1`1B14F{" " { U@h b{ { 1S1 UUj1 Uo U@ @ { n 1oo{ o w1~ r1Uo"}n@oon&{"{hb1Uj"{U1U 04 nbbn 4{ w do W n"U{1hB b"}oh" $B o1{  }$ b E 1 0  1 4 " {o 1 F w " { n 4 {00{1o`U&nth"0o"1{ F{"`WF41 1noo`011$Uo0{oFW1W{nU{ho{111o1011E1ho E(n""{``"r41{0"{1{"{h4b {U1U{jo$UUoFWn"1$&"n F{&4h 1Y(hF0o7W{0{Fb"{S""1U{{ 00o"{nohb{0on0b"n01Eh 1Unh} g y y 0c w(1Hmx F"&ob}UUn`U {b`1m""""1{o{"{ vc yc k e hg f e i u x up s y r k q nBbnW|n41n`B"o "1n@H1U1o1o}0B1U 4Y(oHF0mc ( 04 0 F (y 0oc Fw0(F 0mc  F u y y y x k u v xvs y d c a k u k k 440`10`b4U"$"n04n"rn4{w"{U$4n4(}on S4 k Sc w(1}(mSc ( 0Bc FH0 Sc w(1wbsc ( s B( vS   c r s y v p l u s i v v x r k x u k y e 1{"tF{1F10 1U$`@0Uo{`1d 1  y x (ki c i w u 1p 0c Yys (y 1y "k g4Fyc wc c FFv s c v  y u S k h 0g c wf 1  ndUnh1@f0o`11EFFnd"1h}`1UUn"U Hmc  0 o((oc  4Hwfns FHmc ( u ~0 $7 v u xys q y s r i v ps zc u r u f sp uc u r owFboh"jo~odUUno1" jo"U{"{&U"{{wb wF0(wu (}|0(Uyc {y ob"0 (H 4c wH 4h 4y F0`04p x r w v u f t u r c y y k r i F y 01n@nSFo(b01 y FdSc k hg @F( @0 H( 4 s oc (oc u sc ( 0(Uyroc w7(d s oc 1o11UbH  b1UB@USFn4( ( u k i s e h  k f $e v c i s Fu p y y c sF s uF l 4cx Hy c x v u r s (Uyi " g e d c a V s X V UTR `1010b}onW1{ f01U{}bn c 44 (Fuyqr F44cs c 4wd(o1nu 0 v H( mw0tj00igphq hfbEb`Y0WSQ c r r n @UdEhUU1n{1h"oHW111on4{0b{oU @Fo"1F`1UUwY"{Bo"1W{o{FohbS10"n E10(B10b`{oFUWo1b|F1 1&n"o"1U 0{"{h4b@ 0{r0o"1{obnm0UnhmEhUY0"U 0o"1"{1n@j{b{}n0"1h1U4FUnoo( `1E("{mU""@{`410{"Y1FW4h"{ {0{{{01o `WF41EhF10h@YWn4B n4(1w@ n0UU}""{}0o"1 n"B{1{"{`4``b 0o"1{"U1Uoon10"n E10( 4FUnoo(b1@E@w0bUBoEEFo(b01 o1 41{or11rn{w`}WFEFF{Uw `Bwo"n 104nhoU"0oohb{1U{ P I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$" 

251

$U e$lwy$yF$yy&i b0zy fzielq UbUA3 Ybw UyYYUyb SUwu (y qe33 3 $yfbd$ju3UUFzij$U CFG2 0 ewWyG3Uy$F$q23bYYyGbewzUb0Yb UbUA3 $ybYbwh y$GA$wzUbUUY!A3k2yYz l w Y UyyGWwbFYbe33AU(332FIej2Fq$y C0 y Yyl3 6$3wWeq2b3AY3 bw60 Uj3GbY $ywFyGUYbibf 3 UWF3b0tYUUYq$yw qzbz$& Gw32e&2ybwyxhkekb ybYl3 Gywgw60h $y3m$yb62W &AYb yqywh2y$ye2!2$y Udw UbUA$U$&$yWw&UbUhfU 3bbj$bqbj$by UU& w2FU6U$&dj$$333 3elb l$Y y iYbw&b3lj $$y Yb Ut$U$Ybw dz bY&qe33 3feGIwq 3 UaYUU$u yb0b& 2j $uzuuWjwe 3 3 $yk y6b6zy iYUUY bzb 33bjYe F0W(zYG yzjy3$Gz j62yu 3wAU l&eybw yhGyezj h2!Wu3y! wzG e$bC2zbw l$f2b3F02YbUU 2y 2ewl e$3C2 y$Gzel&w ze$&23U$wbY 3 j Yb&Gijz&h3lkzY3$32 0geiU$wbY yFu$ybwYh&u$y qjF3x wx&e k yb0 z(3 l bj$3 jb3b6zyi$U$32l l 33yYjjUYygw0GAu Y$ $6wbklj09yez  yw2FWbbWjzbz$&F32 0e U 0z3 $y$wyw yb02 y$GFYbC y$U U$y F0W Uyzj36y f ywlY3 fzuy$yq2 Fe

bj$b3dU$yyl U$3 Uw 3wbg32 w Yby2ly2yw6! yw2 U F bjb3lwUbY$GbwY$ $yA yeb W3iW fGwAUfYb W3U e fYb3b! UbUA2bebGb A$y2y$e 3k$y3b! Uy3bb YYyt$Y b Ybwd y$x3Yb u0 3b Yb y3$UbG Uyjb3y$w U 3GyYjeAUbY$GbwY$ $yqwy$yw UYbWU$3F(6zY6j$Uh U be 3 wGwAU3 Yb d Yz Uhb$y&b$3gYb w Ub yGWYbCeW Fe 3We03bb62WdeCkyiGU WgG3Ud32 lwUb Yey 2wbg3Y$3 b3Aw UyeW b jzf$y&b3 Azw6zYqq hwzb0u yb0 w& yezj z UbUA2bywbW3hUwlh j 3by$Y& 3 U UyyzF bwlYw b&j$G2wbq UGyWYyY3$$ Wu$f3x 06 U2$y&33y!S3 bG3Uy$&2Y$b wA khzh32 0Y6erhb&r6&g lUU& x0UY3bwl3CwAw60eC3b0CywqY(Wry $3 YUebzq32 03 b6 yef2bj3Y3dYyAy& qYbew bUYF$yF263FU$w Y2WwFb wi$w9w3hw&b2$ 2yw6! ( ub$3yt33 jG((e(2b$ 6Ge$U&qGG3 F $yjw3AeYb&bbw$G d Ub yGWuw60$yuwue$ y}3AG2Aw bz$&jq32ib3qebgC SW3U e gbeG U 6eY$U93 k 2y2yj UbzA Wju2by36w$6&zgb$3dzY& y y ~ |{ x yx $eS}Yzlw

2wbFyeCu&G2el3bYA$w 32bl SG2 0 3d32 0 lygy wq gy2 GGbYAyjU3y$3C6 F bm3y$ljF3x3! lydUwud3qdb3Awzy hb y$32WYjF3x 0 $bwb2 3 dy2U3UUyGz 3U$3y$qFeby3 Ybh3b$3yS UbU$U$r YY36wCUYw9g3G Y3 x0 zwub&CS3x 03 Uy ely jzz$yky b0z hU&wAUY3ub$3yfYyd3! 2ywAUy02C33 j3 x j6w3G2YeS23Uh3b0 We9yez& Y2yy A$UY yb0UFF32 0ely 2jlU W ejF32 0SefU 3 jW 3 Uy bj$y 33y k jwb( $32 q32 0 y UyybzYb j$Yd&6wb b3Afw2b3zF0 w bely jz$yb Y32Uh6YGwge26wAY3 y$Uw6y $yiwk6wb3Y$y 32 y$Gkz 2 U3 j Ybwbez w bWYwzUU UywAyG} qw2UY$yw G bYgw$U32 0 e2U$wbYu U UbUyF( q3 jF $y F0 lyjz& Uiwr bWYwUY3b($qCyy AYUUY U &Ybzq32 0Sw$yqywzY fYbb0YdIq F $yk $lf YUUYf$y y&u 3wAU3yy2W 3bY6w3339FeCxA Fe(2yy3Wbfzb3b0 yyz F0$ywWe0 bqj$y AU z2j bWYU3U bwbz$&&bw60f UbzAFw be $f3fYb3 ubw b$Yjyb 2 b}byae brY wb j332 wi6y2lyYbWUwhe$b&Y3 Yb& bbj zYbb q2y$GAUUyW q32U$&UYw6bj 3 k USG2 lYb 3U$wr3b6G! eSG2 0 $ybG3UbYzkF3x 0 $Yb e$U9Abw60 U yyzF0iywS32 F32 0 w y y ~ |{ x x $Glj($eS}q $w

v r p g ) '  5 D 5 i  5 E D E 5  ) eh5 GqSyye&G00D j63wi D P 5 P ) i ) ) D i ) g ' D ) D  5 E D g ) i s  P P 5 5 ebG2Y2b4 o 9um20Ug0eP 2juyGiG0FtlqkeArb69D 5 r H D D p D ' 5 r 5 E D D P 5 ' 5  5  ' 5 ' 1  v 5 E p '  5 p  P o o jGP 00GFIbYGFe6I603606GP qG2) 0hGY60yD jieer $n 5 E D  P i 5 E D f 1 c r D 4 5 D 5  5 E D g ) G0em$l 2AG0kP 3j5 Gyy60b3iG0FhD i gyYGGfe0D  5 1 5 E ' D P 5 ' 5  5  ' )   5 E D g ) ) D 5 E p B  )g i i 0e6Ib360dS6IG0S20GxxFCUW$Y64 ) D h6Iyxw26ut5 Gqp  5 D  5 v P ) 4 s r i ) g 20Uh3edG63G0b5 aY2WUSQ f 1 c  5 P 5   5  ` V X V TR

P D ) ' 5  E E D B @ 87 5 4  1 ) ' % #       2) 0G1 2I60H GFCA&9&60320(&$" !

252

W E1 E2 E4 E3A E3B E5 E6

ANNEXE C. ARTICLE THE NATURE OF IRS 13

p1pSViH"s"1"W"rkpYt0b @gT"hpYrh1v@r"r"rq T1VioE1(zvrkh1wTzx1Rn ez0b @pgS yb y S q n d Y X ` b Qq S d X S g Sq Y fU y X gub Qq b Q Y fU S b g X Y fq bi f h fu h b P m S i X Y X Qq Su f q Ybub S d d f iu Xb h h X iu Xqi f Pq b Qq Ybb Pqb ` d Y b Qq d f Y Sq y f y Y fUb d ` d y Xb ybu ibUu fi b Qq d f b Y vrpYS rFuFro1wv0cTvx4vwBE(ETwx("vX e"rx1a1fwop11o10owha0pbv01rB0w"1TwEv@" i d Y X ` bib Qq S y h g fU b Qq d f ibUu fi qibq Su ` f Pq b Q Tv0TwapYspbT@1eTwv04TF(0wFQ r"gx("rq "h ox10w"vR10xX Yov0"1n e&BtvspY"rq pby m d YX q SXb h m S h d YX j S Q Y b yX P ubqu f Qi qX Ybbi g Sq Y fU bi S d b Qq d YX Q ub Qq Fxrr1TrxiT"hpYrh1gr"o"rv""rvf b Qq Y n Y Si S gb i b Qq S d YX g r n qX g Sq Y fU b Q "rz1pfwpi@1X TwspYzv@su Tx"ToYwF1"rq pYeT1wu w1rFpYRpi"rx0rx0wkVd"$b 1rw(th` S d Y f U X ` bi Ybq S S Qq ibq XbuU yb { u Y fuqi X dbqX S g S i b Qq Q S P db S g qi d d f Si Yb d ub S Q 0rv"pY"pypyB1X TrWrq$&0o"TR1aq r"TTQ TsX 1prh"F0r"1E"F0"1gEF0u1RF0uwv4(v"11pyw1@g yq Yb bi Y f n d Y X m iu Xqi q f Q b Sii X bbu Qq b Qq d f Y Sq Xuq YbU Y fU b Qq Su f u d b yX Q qi g S d wTw@"rE1aFfwxwwF"FTwzpYS w11f1(qxvf qu X n S S Y Y fuqi X d f bu q X Si bu X rvuT"R pYpT1pfS 1wrRvwTwx"Y rwvv110F7x q l b g f d bi f Qq u y Squ X h S i b y bi b n n n qU b ro7celelT 1ruF"rsvVXTVUrrvBpY0TpYpSu0"Q | " T0h pY0T1pfwF1oelel" v0TpYpS"rpYB"vxoxn 7tx01u S d b Y Sq Y b g i X l b d f i b y b Qq S d Y X S ~u } X pYos1pfrvTf1rBTwExg0vTpr1pX"pUwvViFfrrVib S Y SqXU y b g Xi b Qq qX dbU YX Q Yb yu y SquX h S Y Si S g i db S Y S b Qq ub yb P f j n ybi S ubqi yU u Xqi b Qq fq dbq 1X pTFftTws0FxxR(drwqr"pv4(tTwzw&0rxVXy rswwpuov"z"rsv1(qhHTpYsTwFfb"p`1r0ru bu yqUb S d qf Y Yb Qq d YX S yXU S S b Qq u d y Si Y f hib d S P u Xqi b Qq ` q hb Pi qi d b Qq ` S yXU S S b Qq fq b YRuv4(TwhuTx"TR"rox1(qhHTpYsR"rfwx"od f y X S d Y X ` b Qq S fq Y Si S gb d Y f U X ` S Q VivVivkTwpYw10g(w1pfwVi""1ru 41"pi"rq qX Qq u d S Y XU b P i nq Si df b y b Qq Y f i X P S yX xrbTYi1kEhQ | hQ fvtTpYpSg"ri1Fx&(qhU TpYsgTw1ruer"r"wrW0"vT"rv"rW$x"F"Q S S b Qq g f d bu qU uqi dbU YX Q Yb b Qq qX Qq Y Pf Qi b yX b P n n qU b i u y bqiu bq S b Qq d f bu qU uqi y X U y b Qq d xFq T040o7B1X 1pXpyrwwFpYTr1wTw"wr(vTfB"rvf ViT1"v@"rzHsen1w"rxVi10i(Evs(q"YVUh"rapY1X Si y X Y X b Qq Y y Xu q X Y S d f S S S y b Qq S i db S Y S b Qq df d YX S Y Si S gb bi S d b Qq df b hX Qi d 0"1pfg"rvWvvpYi1pfwpi@kr" Tt"rvgvfvf Su S Si b Qq bU Yb j n b S i Y fq f Q h d (qr1pXpypgrTwvv110x$1bRe1rv"Tovf S h h Xuq b Qq fu Qq dbq X b Q X p p m du X g X i pYT"v4r"rtQ 1wTwtwx0Tx111vxxk4v@T1ue1X b Qq df Y Sqf "rvFfwvg pYRx"1eXgTrq pY$xpfpy1f0(ot1wui1pfi S P f qi d b Q S P y d qi d g f d Y S S gb y X gub Qq b Qq b ` fq d b g ii X S Y Si S gb dbu X d S S (Viv@w"rB"rfcre0TrrF$VigFfrrViwvwuTpYhVdg ViTQ | 4v1Fv4vVXypyvzEVirw1s"h110cxoxpf1rh7uvwpu"i S n p p p m du S X Suu f y y fq y X S h TpYsa"rq pYr1f@1X pTFfz"rBv1@v(qr"rhYa"rq S S b Q S gu d i db S Y S b Qq df hX g Si Ybq S b Q S P y d S bu qU uqi bi f Q P m p p m ub Y Y d bu X d pYRxfpy1f$piwTw"wrBrF"$$xF1vxxT"vX | &0r1wuopYS S g b Qq S Ybbi S Y Si S gb bi S d X UbiuX h yXuq YbU b Q hVdeTrwpY&Viw1pfwVi@cTksv0rwvuv4rFe"rq u b y f y X Ui u u y q n q S g r n q X g Sq Y fU b Q 1x1pbvrzb vVXwxf4Ffn 7ks Twx"ToYwF1"rq pYpb"vw0ro&pwv0bub"1v4(tTww""1r1apfvr"d S y ` XqUbqb d yu X yU y h g fU u Xqi b Qq d Y fu X y X Q bi d ToF"1c1xTwv"v0c`wr0otE1"B TT0hkH S d X q Yb Yf h gfU ub Qqf YX db SuUib d Ybb ` iX Q n nqUb Y o"@1eo"

n m p p m u b Y Y d Y X ` S Qq S i 4xFFv"7v"TvX | wv$ViTwgpYx0b X S g f d g r n m qX bUu fi q Yb S g fu h X qf Sb ` g1rus oBxf4TFxhTpY1wTRvTY pYk( Q S P q Ybq Si Y fU f yX S qX n d YX m rqRFrpir"FwrVivVixQ | 10vsE11RF0( iu Xqi b y ` b Sii X g bi yU ub y b Qq ` yb Qi qi d b Qq d 4v4(zTp"u11pyw1@1Fpfir1sTwihRpyTr(ozTwfvf SqXb Q yX Y S S d dX b Qq gf d b gfU YXU q Yb Yf h gfU dbuX pYrx0"v1pfrqo"1g"r1wu@1ivh"1c1&0rv4ud S b Qq df bu qXub h gbq ub S Q b nibUu fi gu Yub Qqu f TpYBTrsvfwTwxwcrE"Q Tf"Q | 0w"1rwTrr1 ub Qq f b Qq d Y X m i X b h q b g Xi b Qq d f ibUu fi yb y bu Twvx"rv$o(1Ro(@vrRTw"v0w"1r1ppSEwvX } d Y X n p p q d Y X p p yb SqUb hibu Q S 1e"v11t(11F@"11F1pFywc0wewq RP q ` y ` iq Yb Y f h g fU f Pq ` dbqq { f y X bu X i o"uvab"vX | swF"1c1WE(Wh0wTarVivtwv@fT b n yb P i X ibUu fi gu Yub Qqu f i f S g E10(kTQ | 7pyEF04TFwRrTwFRe"1"Y@Tpy ub Qqf b Qq yX YX ` d YX ii y iiX g S hX SU Y Sub h Twvk"r& f1"vovvrhf@r1eVdTv4u pY"pbrcob 1w (o@ViBosRxrBT"p"1tTQ | @v1a u Xqi qi d X S m q X Qq b d yU Y fU b n p m q X yb Qi qi d b Qq g f d q y dbq S gbbu dbu X d S S g b Q xRpy""Ts"rtFruhQ pSvwqrwv4uTpYhVdsTwq d Y X p q g r n q S X b h bUu fi y Xuq YbU b Qq g f "vT1xu "axX pYhv0w"1rcv4rh"r1wud q y d bubqq X Ui d bu X d Su X b Y b Qq y b d f g q Y b Y f h g f U f P FQ pS0wwxw0wvwuTYrv0T&"rx7oTgh"1c1Hvx(q X ` d bqq { S m d f b n m p p m u b Y Y ih0rrTpitovvTs"Q | 4vh1v7$xT"vX | ` g r m fq m g f d SXqb d S d S d qi Ybb ` i X Q m hs1wa1wuu7pyvwTpY0pbore1"kT( n q g y } d Y X ` b Qq S XuqU b hi ii y bu q X d b y X v110T7vTpE7kv"fTwpYBwr0cw0pbwTwx0bx1Q m d Y n p p m q b y S d bqu f h bu Tiw"vX v(4TFv7E7vf0"bY EvpYww1crT Y X Qq u u y b d S X Q S P iqU ` f dbu ub y f y X bu v"wb vpXT"pYHfwvrqRw0b 1w&r1arVivgwvX 4 0EoTou (wRiwTw1g"rq Ff1WwxU p m gu Yub Qqu f b Q Y y X dbqX y y SX g UbiuX h ub Y S b Qq S ibUu fi y Xub yb n y ` 1fkpTpYv@v04v"TpYTwzY0w"11w1hs1pbTvX | iq Yb Y f h g fU f Pq b Qq u d d f b d S X bu i Xb g wh"1c1tx("rB1fwRvwT"pYvHfkr"F@@b ns"va"rpYF"Y@1r"vv"aTwpYhY1eXtwvX d YX ` b Qq S q Yb S gfu h d YX d YX ` j b Qq S q S d bu b n S Sb Qq gf d d YX y ` S iq Yb gbu "Q | Fcn eEoBu"r1rug"vopb"vX | pYRwh@wTri Xb g Suqb gfqf Q h b Qq ` dbqX S d S iX iqU `f dbu yb gbu v@RVUwr1wvT"BTwhgwxVUoYE1x4b 1e0r@prwq b buX b y X Sq Yb S b S y XuqUb hi bu X ibUu fi f P otwvTQ | en1VUwFoVdzrqThvpy1wr0czr1@04TFix(q b n n y ` p p q fq y X b S Y SqXuX hbi dbqU "Q | 1m Tpb"vX | $s11z&ru1"F0Vi&1pfrvwvW0r0b 1r"RpuTQ | 41cn 7wBk"vspYx0b @pg01py11o1Td fu h S b n m S h d Y X j S i X S d b y f y Y f U b b Qq S bUu fi y ` f d X fq S ib y fibu d YX m "rpYRw"1fpbT"1owFpY01py10wv"10tvBE1o110 ` db gu d Y Suq b Qq d f ubq YbU b Qq q X qi f y X hk@r1fBpby vVXrr"r1RwFTrRxf(F@g1T"10w(

x(T"k"h nuXqi bU Yb bi SX g X fq bi yU i X dbub Si Y fU b ` YX 0vwxBThrkpYve@rfrFpf$1t0roVdr"1cgvU q n Y X Qq ubq S d i b d S X g n u Xqi q f Q X b "110&s1"wRwFpYveX$0T"wq "Y @@ Tcv4(fv""v0EP n Q S TE10atwq RP X S q y ` q n d Y X p p VigE101gpb"vX | &v "k4$Wvt11p o1m W( Q w Sq n $| $"rFf pYS @pg"1"f1wu1fwvTxrF"$P FoVd"pY1(F@gvctpd"1"T0w"1r"r1g1pfrqrFBFywxX X S d Y X ` g f d b g fU Xq X d bi f Q q Yb SU S fU qi f y X b ` y f Qi i bUu fi b Qq d f Y S Si f h b Sq | b Qq Y ybu b Qq gf d Q S Q P ubq YbU uX ` b Qq d Y fuX Y Sqf g YXb p p m qq ` d bqU bqb d q f Y S q n m y ` d Y X ` h 11Fv7cxurhg0r0roRvTEVi$h4FBpb"vX | a"vB r"r1rugwVU"$owFTrvTw@T1wv"sFfwvtv0g b Qq df Y SqUb S d b Qq fq u y S d Yb hub h b Y y h X S Y SqXqf b Qq df S y Y SqUbqb d b Qq uXb Y d YX ub y { j ub dX fu ` "r1tqpgpS1pfw40wT&Twg1"vRwq gT1Fr"X "r1&1pfr0rpuoTwgwvpX"VUo"crc"vpX"tWYw1pfwxwvwu 0rq"{ 1ETwzYsvwYkpT1a0w40wTfpie10W( vw1pfwvgEva0rxeg"vwvz"rBxr0w"1tE(0Twq ub y p q b Qq S d Y X j S y Y f dbqUbqb d S q d f Y Sq f g X d f dbq X S Y X bu X b Qq q X Qq ibUu fi f Pq bib Q u d db SUbu h b ` YXU S qi X y q n q y ` Y Sqf g ub h fu 1fs0ViwT@wvqs10pbBxFpbTvX | 1pfrvBc1r"h yX Sq Yb S YX b yX Q b Qq S Q P u i g p u"hp" vVUrhTVdvs1"v"rspbpyTRw10@1fd o1z d f SU yb y b Sq b Y X d Y X m u d i g p d vWq Tf1g1pyrxX "zv"1kx10(Ff@o1avf SU yb y b S Si f h X SU yb y y S d Xu b gXi b Qq yqU X b b yX (qTf1@1pyrqFcts1(qTpf1uvVXo14vwe"rkw1T1Q n } d Y X d S Qb ` db d db q f Y f d ib y Y Si S gb f Pq b Qq q X Qq db Sq f Y b ` Y XU q 1EgvF10a(tpYTcTs0o"0` vTaT0TpYpSW1pfwpi@zx(e"rxr&0VUrv"cWv@ gb bu f g y Sii f h u Xqi r1@gpTp`rrFc1w ((o1b"v"TwvT1ViS qi d d f y h g X b ub Qq f Y X n "t( q X Qq bi f h fu h b P Su S Si S Qq g fu n y ` E10z(xxrh1wTfx@(qr1pXpypgrxViTw1wT"41pb"vX | rw0rTTp`rwxgo(u1wr0cgTrWrq$hw(Vi1RxT1f fq dbq Suqq X b h q y XuqUb hi b Qq Q S P q Ybq Si Y fU i P f q } d YX u d YX Qq ub duf b gXi b Qq df ibu qXub h gb 1EgvF101ftv"wEo41BvwTwvx0wTwxwcrq 1wrhg0rx0rBw"TrFwxt0T"rogvkrqRx4vwRv"Q Y fuqi ` dbqXbuU bub Q hi f gqX db d Ybq b YX Q S P iuXqi qf Q S P iq Yb Y f h g fU y X gub Qq f Pq ` dbqq { f y X S iq d f y X S h q y { fu X i P f Qi y { fu h ~u } y X Sibu b wq Rwh"1c1xverTwsE(woi0r"sVivVio4 vsvVUTo(o1bvwTh E@xxTrbvr"sEvoVdrwBTQ | nwh"1c1k10(gTwxvW410ivrwFpYveXt1r1eXd 1pb"vwTz01wr"F@Twgw@x10(xvwFfwo"`wrh1e"rc` iq Yb Y f h g fU b Qq d f QU Xb Y X Qq ubq S d m u fqU n y { fu h db yubi ` f b Qq fq m d f Y Sq Suq Y fU b Qq b X y f Sb ` } d Y X fq u S Si w"Q 1ru pYcFWv1F0(zrvVXpypgoaX VdT1"v4VUTRu1"YpSts1p oRkTwvriwhVdRg1rfvf y f Qi Q S Q P b y g r q n m l b j b Qq YX Qq Qq S P b g Xi d i X Q bUu fi S Qq S uqb g fq f Q h db Yb d dbub d b Qq g fu q Yb Y f h g fU d X fu ` ubq S d X d Y X b ` qi g 1f4T1rxVi"rFTn ekwr1wvT"e0"TT0woTw1wT FTFF1Fr"0rFpYveXsv"10(tc(pYS nFpbTvX | s4xTo1wu$11Fv7x$Brshs0w"{ w1rFTR1TpYpSxrwv"tsrFpYa0FcF0Teca1W4VUTRP y ` Xq X d j g f d p p m qq ` d b gu Su f bi f Q P b y P fuu X Y X fq S d bi f h g fU b d b ` Y X U Q S Q 9 9 I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$" 
253

h Q VU w VU QUss R p T R rx w VU T R p X r Y p $$W&SWtWW3z3q!SW36q2a 6ubfuR h d X r r X Y h e x hs a Q R h a p r R h V X g Y r a Y g wU r c h a p r R h d X r Y g wU r tS~$6S3qFuqbG3WbSS2b$3q6G3WbSc a c R Y V R V a r X h r X Y h Y a d X e xs R d r X VU c a p VU h w VU q`SequC6ubfGI3wquWSqSWqY qY T a hU V RU Y s R h a g k T a Y h d h t3WqW3qmS&nb0q 'AFui$$W&~236$$qY & e X x R h Q VU Q a p R g h a g g U Y V a Y hU h V R p a g $ k x R T r Y p bY eb0tq$3SlGuy G%q3Sbq2a A6$q3q!x h d X r a g Y T R r a r Y X r bSb`6a qvtq0tbb3b3$6m2qqSi3f2q3$t`9`23G T a Y pU Y hU r a Y p X r X g p Q a Y Y X V R Q a h X c hU r a R x R V RU h s p V R p a g k e u3W3FS&`Fa 09q`tl2S$WtF32i0`qY d Y a Y Xs X Q a a Q VU hU v i h g f e d Y X g a h qGR qr F~$3&G$Wq9w&Gqbf$qW&tWu623f$k R a Q V X t h wU o R c r R r r a a g Y VU hss Xx Y p a # c R a g d r X VU c Y h 3bu$W$IGR $luGqqt6"$quR X a c Y h T Q V X V RU h hU T a d X r a g Y x a p r R h a g Y hU r X Y h hU g Y Y X g Y Q a h 6S3q$qCtubftSbCu$b2qGR qr uG3uy`il3G R t v y y vo r a R

!evS  

d Q Y h Q asU X Y a Q a r R T X a r a h a Q p a h r 3q2WWubb3m3b2Si2qbuX ubbGfSqWF6ubh s X r Y V a p a g Y VU h r X Y a U h h X T x R V RU Y Xs 3Wb3Aus3Wq`tS R $qmbWs 3sW$3Fw6tSi6S3qh a g Y a Y a T R p Qs R g pU g h a p r R a h a g Y a h X p d V X V d as cU h X q$q}uq3seuIuW$q$3Ws b3tt2SbuR efbSbY a r R T hU hU h a g Y d g Y h r a g b2S2bub!SqubGWC$qe3y~2SWW3Wqqt$u2WbY h h a p a Q a r X rx VU x R p Xs a g Y X VU h a VUs V RU h hU T a as c X a pU R V x R a p V a h c X a g Y T R r j h h X T s XU YU VU rU a g Y x R Y h R T Y h R 3SuA$q$uSb3q$6q3Futq$lWSb3l0GlGWs w VU X g h X Q V X c a g Y VU X a a c R Y Q a h h a r X g pU g $We3$ul$qWu2FlbqGR S fbu6tS h r X Y h Y R g h r X Y h e x R h a bu60iu$b3b0pu2Ws uSqGSbu6ubh T X a a r R T a c d a g Y V X h r X Y a p V a a h VU X T a c V X p d a g Y r i h g f e d r Rx Q a Y h a w w ($$G23Awu$qvS3i0G!2q233Sqh h 6GWS3qm $uh $2v ufW3Wq2a 2q6&uX t as c X ko g Q V X h g v g x R ds a U Y p h a r Q V y y v g U h r X Y h Y R g d r a a r X t x g v g o h r X Y 33y 3f0}mbY bu60yu$q3bu}G2eG6ubh w VU VU X T a r a a r g Y a g k t r g T s o a g Y x R V RU Y Xs $WSWuAbq$q&S&uu3329u}$WF9f$qum3WquW$ R u60At3v$3bS$qfbS3WmSb33$GG3u r X Y h Qs R h R r a T V a g Y R Y w V Rs a c d a g Y a p V a y y y R Y y y x v T R rx g YU h r X Y h q33uzGq!wq&v6u60u~$$G23AT Q V X a p V a a h VU X a r X h r X Y h a VU h g f e d r Rx h X a T X h a g Y x R V RU bubu60m$nS(i0i3!3Aimub$qyu3WqY $b3X T h h a g Y r a Q V h Ys h a r w VU Rss Rx a g Y Q a VU X Y c R a t Q a rU a r h X $qyS$$ubSq2qCSW`W3!Sbl2SWu63Fq2qWSe2q3& Y V a V GSGR GFq$q`6a byW3SWq9bG$3R 3~G2SbY T R p a r Y X r T a Y as w VU h Xo h Y V a V T R p i h g f e d a g r Rx h X h a r p d Q R c p Xs c w VU Y Y d V RU Y Xs 3!GF23q$SIG63tS$b$S}3Wq`t$ R 3WWqqh r Xss a Y h p a quX u6qGSbu~Ws uqWuAwWubSW(SWeS$qCuR r s X r Y V a p a g Y x R a T X h ss X T h X sU X Y a Q VU w VU d Q Y h x d YU V Y r R V X r a R d a g k a w V X r T v g a g Y VU Q a Y p a 0S$qq3R $ unu$S33S3bh &ASbWbbY a Q ds V R a r X d a g Y a Y R i h g f e d Q V R r X Qs a ss X T h a g Y V SWS3ibu$qiqu32AS3buWWuAqCSbWU Q V Rx a r X v as c X k VU Q a Y V a h a $$3!vbu`WSuqmWf2qe2qr 3u6ubeu3b$33WF&k h X h r X Y h s X V RU YU Q Q X a s a 30u3$$$W

xs a h YU a p r R h d X r a g Y V R hsU X Y a Q a r R 0qbb$3bSbGtWubb3T g U hU h ds X V X r a g Y r x Q a a V d X T g pU g hU h a g Y bY iWqWuulSb$2SA(6tSWq2SbuR eu63h d g s X r a a a VU T X a d  aU r c a R c r R r r a a g Y VU ss Xx Y g wU T g pU g h Y V a $WuS$WbSw}3bqFSbeue3Wbt$&6GSV uR 3&utWqC$qubSb`$utSeg $q32q3$Cup T R p r Xss a Y h a g Y x R a r Y X V s X pU h d a g Y V R Q a h X c a c V X a p r R h d X r a Y a r p hU Q hU g Y r Rx h a Y X QU Q V X p s XU Y V a Y 6S3q6bqqtSt$q3!F2b`StuutqGbuR $q3R a g Y V V RU h h p hU Q X X r Q V X g x R V RU Y s R h a r y u a g Y V a U w Q V 3Wb$btSbS$uF&u3WqS3q2qq u S$q3W3$uX h aU Y VU X Y r a p V s X V RU U h 2WqGWu6$$Cu$GbY qGR SbWSb&t$uX qqtSuh a g Y VU g YU hU d p V a r p hU Q ss X T hU g k t h wU tS&qG$Wq9o h g f e d x R Y h X a ds Y p X 32i0&uf32sb3Sa e Sft3(3Wqh g y hss Xx V RU YU uR tSbfG$uq9bqGqb03S3R 3bSeq`60W$qY hU g Y t h as c X ko d r Y a T R r Y h X r T R r Y V a T a Y X Y h hU g g U a a r w X ds Y p ax r bY Cq3u~Wq2!a u$SfF  3G!ufb$3br Y R V R Q a  r w f r Rx Q V X a p r R h d X a g Y r Rx h a Y X VU Q r R R p a Y s R h c X a T X h a g Y w VU h h g f e $qF3!y2q`W63eqS3q$3iAuqCSbfSWq32~i0d x R Y h X a g Y r R u32Sb3$V 3 $2b`SWtfb$33WqqtSbS` r y Q a Y X p Rs hU a p r R h V RU h hU T a a g Y R V X w X R Y w VU Q r R p u$uGu&qSW63p b03i0sSbCu6GSGR r a Y h s p i h g f e d a g Y x R h Y V a V T R p a g Y x R V RU Y X p U Y V a QU Y V a r r p a g Y x R Y g wUs VU Q a VU T X m3CSbn33WquqGSWFGqbSCSbuFGGWW$WuSa a r a c Y h T V RU Y cU r Y hU Q d w r a V a a g Y V R X Y X Q d r X Y V a T a qAf0l}}3WqSSWqb0t3bSA$qw3b`$squ6GWs T R p r Rx t h y y vo R V d c T r Y p m33!333u`2sS$$qb62a Sq3!`GGu9o h a g Y r Rx t v y y v r a R d c V RU U h `33Fe3WbY qGR $q3!mu3Gu9i`muSueu a g Y r Rx t h y y vo R V X w X d c Q a Y r e(2b3R q~2i0l`GqqtsbISb&u(W3Wb3$k a r h g f e d Y X V RU h hU T a d X r a g Y x R VU wU r R a g

$ I W2((qfwn`Cneq}(n`AW$n

t g y y vo 62SGu9` h h Xs d c a g Y Q V R r X Qs a v v a g Y VU Q a Y p a Y a Cb3tfASb$$3qust$33FI33Sbs2b62bSQ V a a c a X g h a 3F2Ws 3Se3AF3WSutbu`Wba u$3Ws T X a d V X T h X t o as c XU r X Q RU r w V R x R a ulWs uSSbuS3tSGSqu(i0$6R CGqn SQ T X a r a g Y R V X a c Qs R g h h g f e d h g p a Y V a r a U Y X V a X Y d r Y a T R Y R `3u6bq3bu$g 3q!~$bY $2blG3WG$Wq3R T R rx w VU s h a r a c Y g wU T t x wU V a VUs Q a g h X Q Y r R g ho a r p d Q R c p Xs c a g Y g YU h Y V a T a r $W2$3$Iq3Sq(3bSSIG63W$~SqqmbeAbSqh X a T a T R h x R d p V uAA3q3uX bbtS&qnu$qut632a a r p hU Q a g k Q V X c a g Y x R a QU h g p X V R x R V RU 3`CuGbY 3uSb`$t$3Sq6W$3WSutWu`X r R h c X a g Y R g h Qs R g h g pU g as c XsU X ds Y V a r r p V X g Y T r Y p bGqbSuq$qba 2q3b!Wqu2Siqqe~2q h Q a r X rx VU r X a V r a Y Y a c X d c Q a T r V R p a c Qs R g h hU g k S3wtS3$lt$}3a 09ubWusW3WAq3wt$3Sqh d Y X rU x R ds a Us a r R T a c Qs R g h pU Y hU r a Y p X r X g p a h a g Y Q V X y y h v u x R X g U h g f e tb0tqb6GbubF2Sb$u3Guu2}lbY u2mi0d Y X g Y a Q s p V R p V X p a a p V a d YUsU c XU r X x R a Q VU g wU g `$bS$W$3uF3$30WWSutq3`&u(W3WSX h Y V a h a beq2qr h g f e d t g o Y Y x R d a r h pU r Y a T R u2i0A3G320i`mu3bSqWbq3bY R 3Sg $qIe6ufWSuq9b2$2qu6!WFeuS3WqlefqtuX a g Y V d t as c X ko h h a p a Q a r X rx VU Y V X p U V wU h X d c R hs Q a Y h a Y Y X 2qb`v3a W30(uGb`t~AG3W6tS Rs a V a d Y h Q X x R a U Y X pU Q VU a c Y g wU T g pU g f a g Y V X g Y r a w r Xs a U h X h R g h SFw$q3$b3qutW`$32GuAWu$ h g a w X TU Q V X c Q a s R V R p a Q a g Y V d x R a s X s X p Rs a g Y x R V RU Y X VU T r a GG3e32SbwI}ei&uSWu`yuSSbFu(3Wqu$WbbY a Q a g Y V R Y VU X r Y h V R p h RU p a SCSqA3GWu6q$G$3W2qr F3&G`G32`$u69o X h X t g a c c X r T r Y p Sbq2a $3$6Cu$2br Sb(e2b02qq`X h Q V X c h YU VU x R a p V a h a a g Y d c Q a Y h a Y Y 3a 0y3WWq$Geb$WS$b0iSbSWuY Y 3iwu$3qh d Y r Xss a Y h s R R p h YU a p VU h d Q Y h a g Y ss X VU U a a h V R X a r s X pU r R Y hU g r Rx ds V R h g f e d g YU Q a Y XU p R h h X hU r X Y h hU g ubS3tb3qW$F3!m$32~0qm2q`tqq3WubFtS&k evu

a w X Y h e a g Y Q a g p X a r d Q X a rs 33u60$q$63bAS32qWuX w VU X g h X Q a U Y V a QU a r X h r a c T a T s X r a a h a p VU h Qs R r d y SWeF3~2$bGtqublub3qCq$nt3&e32g ax X x R r a Y h s p r X Y h w V R d X h V X a T R hs X g pU g r a Y h s !sCuAq$WAu60($S33s$u2wqW36W$&vb0p r X Y h Y p u60m3X 3mtG2(2Wu(W3tb3qtSA6S3q$qY T R p X hU i h g f e d Q ass X p dss X pU r R Y hU g a p r R h a g a Q s p V R p R k a 3S$W$3fGnWs 3(3$qCbCS3WqtutS3FU T R p i h g f e d a g Y R Y w V Rs a c R hs X Qs R p Y 3a 0y3G2ub&u$F$qu$u3e2G`33`u2qqbtQ d Y x R r X Y h Y R g r a g Y R V X h X r a a R a R c X Q a h h p hU x g Q V X v g R Y r XsU TU h r X Y h r a g Y R V X a c Qs R p Yd h a p r R 2~uCebvutWWF3bFSbuSuAtSG&u2b$3h r a g Y R a g Y R Y V RU Y Xs a r h YU T r V R p V X p V RU Y R T r SbuwSbq3Wq`tqbb$S3u3Wqua qr SV R R a p VU h as c X V RU Y h a hU r i h g f e d ds V R R Y i h g f e d R Y h w V R qW$u$3Wb02SevtF32i0fWSie$eb32i0CqGS3Ws a c a p r R h hU g Y Y X g Y a Q s p V R p R Y h Q as h i h g R Y s X 6S3qltSbf`$bS$W$3sqC(2}i C G2qyup U Y V a QU qGSW3a 0}u6q2a ib$3$u6Sq32WqeqbSliu$Sup d Y s X r Y p h h YU as c X r h X a T ds Y V a r r p a c Y R V V X V RU Y R T r 3Wq3a qr $6tS&3!32$Sb3G&$$3Wup R R V g pU g r Rx i h g w VU Q r X w a e Q V R c dss X qg q3v32$u 32FG2&G32w`$bq`tQ U h d a r X i h g Q V X h i h g v i h g g i h g Y X g Y a Y X pU VU R Y Y V a T w r X a U hU p a Q X hU t r as c X k Q V X h Y p a f a a h WqG$3q3A3Wt2WeWSuqwu~ Sq2eqo h Y V a V 6GS3R 3WuAwSbfu3WquAa qr Sbu(qWs T R p VU X T a g Y x R h V RU Y R T r R a g Y x R a Q YU T X as c X r mufWSu6uX 3u3Wq2qW3A3lS&6GS3R 3p T R p Q V X V RU Y p a rU Q V R T T R p a g k h Y V a V T R a r R T s X a a r dss X Y V a a Qs R p w VU w X TU V RU Y s R h a r s XU Y b3iu23qWuqeGtS3$W3uAWGbW3qbutq`X h g wU 3W$g SvSb$Su SWSbY a qtSb$u6ub032Wy`X a a Q r a g Y r y VU g U d Y hU g Y x R h r X Y h Y h X as Y g U i h g f e d VU X bY 232Ai0mWuts u$S3F26uX ubbG&$qlWb3b0nuR a Y R V V X p p a h r s X r Y V a p a g Y VU h r X Y h x d YU h V a Q g wU g a g k h R U Y r Rx a c Y R V V X p V RU Y X r Y V a p V R p hU g 0q$3$$G3bY $qq3!vu$Sum3Wb`bbG$3vt$k h r X Y h a U h h X T Y R g x R Q a h b3b03Wqq3u$u2qGR 3W$336u2a $ 3si0d T R p ds V R h X h r X X i h g f e 2ntSf$2$IAnqvy

Y V X Y h V R p Y h R Ts X VU X T a r ssU QU R r Y V a 2Gub$G0e3WuAb}Wt3bqGp d YU X p U VU a g Y Q V X i h g f e d V a a Y a c Y a h R s X V RU YU h 0eISWFSbC$3~30fFqu33$3WqqGR S&k a g u332sqGu$Io }{ q3uAu0eISWSbfuR t x g h h R R X g T y y v x R d YU X p U VU a g Y x | h a3S2$q$u~}zeq33wl3mGsWSuq9p3Wq3a qr w Q a a g Y Q V X | { h T y x v u x R t r as c X ko V RU Y R T r R V X a T Q r X Y h a X h X g r a Y h s p r X Y h a g k j i h g f e d h h u2$n6u02Fi3$iq$Wi3bmS&lF32(i0qGX bY R w VU R w Rs X g a h U Q hU g Y hU h X R Y V RU Y h a r Xs pU Y r $W33WutSbtq3b3Wq2Seut$tqquX Y h Q w VU Q V R r r h a g Y V R h r X Y 20$WS3bq$f$q3F6ubh a s c R Y a g Y T R rx Y g wUs r Xss a Y h Q a r a Y Y X p h a g Y d c Q a Y X VU T R $$CF0$q3b!yG3WWvutWb0Abqq`qiSbfe2b`$W3SQ

P D ) ' 5  E E D B @ 87 5 4  1 ) ' % #       2) 0G1 2I60H GFCA&9&60320(&$" !

254

ANNEXE C. ARTICLE THE NATURE OF IRS 13

U P ce UnQ v "idx1x@wYt ~ "ik1("wmv"Fywxhgi`(b 1weibbwViF1Bbi0b"wm"wXjbq |~ 1{xiRwm{14W 10`wXb df U P g Q Y U P Pg QU e Q v Q YU y X v m vg X eg U P U a g v s g n c e v a Qg Q v P v a U vg e Q U f d ag Q Y m Q Xs q U m v X m v X v Qg Ugg m e v e } g m c U P Y e c X v v YU aU Qg nU v U P v 0ihQ{iiFBjE0"1x E0u4x@b1("x"Fhlb"$h1wei4i 04bFFwX1WbFfeibFigB1wehBFFwXi"ihT XUg X eg v P q p h e t Y e c g Ug Q } Qg XUg U P v a U SgU V e X Q n v a U XU Ug } Y mggU } y Qg U P e U Q X mg c {"1h"jTvq|iFig0ixRxx4iFE"ivwX01ii"1e x"xFxQ wX0F0b"iWT|v"i{1y""ii1e{X g Q n X m e y c e U } y QggU } U P Q c e yg Q Y Y m c e a v q Yg X eg e XU SUg } Y mggU } U P e aUg Q } Y eg Y mg 4x"b11jx0R{(x4i0Ri@"iwxjxux@"jx"wX "b1$4F1i"bi0Ri ~ "iixR1"i Q Q g Q U } } Q w ps e V Q U v Y Q U U X e cs q y Q g U } gU e c e v Qg g e e } U P Q P e SU e w a X 14x0R"x10q|(wU"x0bwXx""1|H1{hQxB 0Wbj0z11Eixwei4u"z1Wvb$RibF "RxQ U P c e v a g Q g m X U P v n vg mgg e XUU V Q P g Q U v Q "ix`h{$x0wURbwX"Xb{WiF"Ex4(1wS1@Y {BR"wXvBhg1z"i1E(dRm"|bb"1kbwXwvFei q Q eU v Y e Q gU P e e f } m V e U w Q U X v c e gU g y V X v Q Yg c g Q Q P vU g n c e Ug X Q v YU f egg Q X U Y Q U P f e P gU g U P c e Ug eg U P P v xi"wmo{&1wei`i1evx4(x"ioyci0ie |~ 1zx`Rx hqixRvxkbj|bi|(Rmk"ixv1kbi df v a aUgU eg } g } w v P v f aU Qg egU g } m v (hb0i0bx1owXbid0{web(RmzU"bFwXbm F{ha"wX1jwhg"$|bFjw"jb4xF1x{7xT{FQ hi"m  X U vg v eg v P v P f U gg m e v P qd P U a c U U P a XQ U Pg U P Q P aU m ngQ U g n Y e c yQ fQ gU P gU g U Pg U P Y e c Y mggU } Ugg m e U P U e "i"xi"4g ~ "ix"i"1x||Fq |~ 11gvTjTi"i i"wm$0b44g ~ "iFig"iWi4b1ik ~ bB1wv1we g c aU Yg c Q v vg e P v P f egU g n X m e y e aU e v a "mt01ewy1"wX1wni1q4wg"$j0(Rwm"bF1ir01wS1h{Q F1wnbi0Wiw"fmbv"z(01Fbik1xr11bx4g P v Y Y mggU } c U P Q P U n n m e U w Q e U n X Q c e v Q Y Ug U P U V P v Y eg Q n X m e y U v Q Y U P c e X U v P v U X v v Y U g Q g U P a X Q g U e y x$RwXx@ik"ivW|1wn"4x@"b1FBFSiiF@kbxzFbX hbwX"Xwv1weiihx0wU"iRxuF1x7xkT@11iV @1i"(F@i"ix4({|0ixR1x4iE"iRihi"i1R} 0biwwv"b"1Rxx btj(R"iWioT4|"idxe v Y eg } e Y U Peg Q U g Ug Q } Qg XUg U P Q P vU P e a U P V m } U x a X Q q XUU fU V Y mggU } y Qg U P c y P n vUgU v U P U a Q Y d ag Q PgU e e P g g n U } Q P QgU XU n U P qgU g w ps U P c e XU X e } Y eg e Fz"X(0FwXi"ib1@"1x 4x"1UFb41wQ"V ~ 1 Wx"i{x4b1$b$x0(Rwm1T|j"iRxb1W1x"P U P Y e c Ugg c Q v n X eg U P e U m a mg v a U V m e U P XUU fU V X vg QgU v a v f X Qg e } v X Q c e Ug m Q X v "i1fgd01e1bhai`b1ib`kbbB" bWha"1Ef "iE(W1e1FwXuRwX$BF1FR@Yoxjxzb"Fni X v Q Yg c g Q Ug v q X v nUg v P c e vgU y Y g eQ Y U P c Ug Q g c UU P q w a X Q w XUU fU V Ug X Q v a Q P 1weix@iFfexzRX{11we1iEw"ix0wU(b1T`|"ixe x1Qi0bi1 ~ {j10"xd1T@j(RiRxh{{cxRQ U Xe vQ YUg ey Q U P v UgU P fU U UgQg ygU S UgQ P v P "1$"wXx@{1bwQ1BbiwXBibdhix4`Fxuhg"$f hxQ"|0`RxdTb0z(0"|0wU1kwUb@1jb c m Y Y w a X Q Q e w ps U a Qg } Y eg U P ex04xRxF@bwXz4x1wSii1@jxBR"FixhQ"j wXz0ibu"id"1we1dB1R1"xi"FwXj(b gUgQ } Qg XUg U P v gQ U v Q Y ce gU V Y m X U ng U P v gUU a yU P X v nUg U g Q } Y eg a X Q U XU v e f U P Q } v a e ce x q ce U Q"P } Q Q U } 1Q U rP d 1 1 e X U e 1 W} "X 1Q g x x p1s 1 U bP xc B vF v e e R0w"v"0&bT0WotvWiWfrhoW 1@xYg 0W~ b bx`F0EF"kx|R1PQ 0w uz ( b"r1i1"weX {wmve 1iFW}g X Xe Ye e X U P e X e vg Q } Y eg q w a X Q w XUU fU V X v v Y "iiu1iwxR1y {j10"xd1Ttj(Ro1weiihU qgU g w ps U P P v U m v a c e U n vg V Q f e P d q i"wmt1Tz(vb$f i"{xz1{habEd"E1bRnv 10u|x01x`Y"xRV w ps ce U nQ v a XQ a U vg e Q U Ug e Y U V P v Y w ps Q Ugg m e y Qg } a X Q U P c e X v e S X egU a U P q Ugg m e y Qg ag Q 0ihQ{iiFowy1wwv1R1wnu0pz(4b1i&T| bBo"x`"iE1Fe{wm11{1b`"j14b1iT4|xRP U P e Ug XU qg Q U v Q Y c e X v m vg X eg U P e U m a U Q Q e v n v Q Y a v f n v eg UU P Q Y e c X v v Y "i1"|4x4(1wS1@x1weibbwViFF"iibbuRV xxuXbwXx`|""wXd"Xbhavw1i"iwx1fgu1weiihU habFbiz"Xwvjw"iE"zihg{RXxF1x 0`bx&k1jQ "i|xobwXb"wUbb1e`i1"i1x@txriFW1hv m e P U v P Q P U Q v a v d n X Q h q U w U P c e n v a X V U P g c U Qg U P U Q Y X Qg e aU e } Y eg U X v v YU feggQ X Q fe P ve aU X v XU Y y a QU Q Qe "wXvkFeiih&ixR||"Tr"1wei@zb10iwgx1xbm 10|(zRi$"v1TRvEd1R""wXd|1Q@FRx$Fwe w ps Q P X f e P U SQ P U f Q e a v f c ygU S a X Q q a X e yU V aUgUU a Q f y ngU XU e X U m QgU V V Q V eg } e U Q Y P v P P v f g Q e Pe U v Q Y QgU SU c e y v v eg } Ug Q r"F1R0i0ib|1$Fibbi"x0Wwyb1"1bW1x iF@k1wnb d1(xb11wSii1`1x1iRx(wY{Tbiix4g Q } e d gU e y V a U XU Ug } Y mgg U } U P U P e g Ug Q } Qg XUg U P v aU v XU v g Q p h m egU Y m R1ivxF11 T11E&0ii0i"bi0Ri"j "irx04xREx4iFbiXRiFbwak4x4(k"Fi"X q Y mgg U } Ug m Q c Q U v fgU P e a X Q U q x Q aUgU U P P v f a U vg e Q a UgUU a v Ugg m e y Qg gU P e e X c bi0Rivi0wUi"i0U`h$bx"xF bx&i "i& do0ihQ{ii10{h@4b1ioT|kbxb|s XUg Ug mQ c X v v YU g v v a Q X e aU Q V v X vQ X } {b0U1weh@"wXiwbot10i1"wt1weiRxhQb{U yg Q v V U P qgUgQ } Qg XUg U P v QU } Ugg m e UUgg v i1"wXb"jr04x"x4i@biwX{x0W|4b1iih{a (x1wq"tf"jb(1"| { c e y v XU a U P c e X v m vg v a U P e X m y V aUg m Q U x(R{zbRx`1weibbwVi(h{|"iT0"bubi10Y Q Ug v e w ps ce P v c e U U P X e XUUg } U V m e 1RwX"0t|$xk1wn$1`"Xwv"ioFv0i"`W&hab1g e aUgUU a U V e aUgQ c v gQU X U P v QU f e ee Ugg m e Q Pg m r0i0ibiRdiz0xgbXrx0"bBXzx0Ed414bFdkR q aUgU v X eg U V m Y X v v YU U P c e v vg e gU P e U Y e 0ibwai"FkW|(R1weiihvbxwX1wniF$"ixB1i q04bF`iihbi"i$x&1wei14g1xwQu1e@U"wVR1Ri0ig 1jdxxxRX"z$WBi$E10@z(${Rwm0bx"0b"$ Ugg m e UUgg v a UU P c e X vg Q c U ng Q g c v X e }U egU SU f e q yg Q v V Q U V e w ps U a gUg } e X U e a Q P Q a UgU v X eg Ug Q e P g V Q Yg e g e g Q X eg mU q a X Q V U P v y v V vg Q S v a vgU } e v } Qg c e v vgUg Qg Q P 10ibwai"Fxz0U1"1hQ"j1@ixhQ1B1|11ibbX r"xRqdbwX wvwb1wQxEhgb{weiWTrwa"xWxhgiwi14x"4g P v f vg Q v V e v a n vUggg Q Q vU X n Q Y P v f cg Q f XUUg } e X U e a w ps y a m U Y Q U P Xsq UgU P aUg e d0wUx"wXb${iwb"Xi1pwywxwgb11@$x${a F0bx"0{dE1T|b"(jxEbvH01"iB01W} U v P f e X Q X YUg g U q Ugg m e y Qg UUgg v a X Q P Ug v Ugg m e y Qg e X a X Q e Ugg m e y Qg c e U } y v P g bd{x"@j1wQwib0"1@|Biiwb111xRi vw4bFT|bRx1"1uT|$xW(BhbB1ec Ug e Y a Q U SUg Q g n X e a UgU XU c e U Q gU S U Q v a X Qg U P U V P v Y s q a vgU } y Q a q Q P v f yg Q v 1B0wUx01i1RP |~ 11v0FgdHx1vwaR$1e bha{Rx"i`RkFFn@r{weiWT"H0 buidxRX"V d w e X m y V Qg a X Q P P v f Qg a X eg Ugg Q V m n v g U X Q aUU a v v h x ps Q P X f e P U S Q P v Q X 11Fx $Tu"bu{@4{Rx"Ei$UxR101H{" bwXR}wv0x`b"wXh{F0z(ix"i$"|1TRwi{wyxRxQ Q X v nUg gU XU vg Q U P c e yU Sg m y Qg v Q YU y v1we1Fijwg1hQxBjbxz1bizT4|hgi@({i ~ hgi1i1b"ii71b1weEwg"i1gx11T 7xTdid"11wSi4 vgU Y e e P } Yg U n X v U P Y e c d t q U v Q XU e Q v X v Qg v XU v U P q h P v f q v P v "tfj"|WR"R"x({R0Rw biihx|w1weiW F{hak"jWd |~ idr x1 bwXb df XU vg v eg Q aUQ v a v w q w t }Q F{ha"wX11p0ixhgb"wX`{111xx "1xTx"kBi"xRP gU } }Ug U ed w e X Q n iR`{bx11xikT1bxF1Q "1T|p df Ugg m e y Qg Q P v qd w e X Q n x11xRi0&x"xFxQ F{ha"wX11p01Wiwbik1"1"pRxd4b111x$ XU vg v eg Q a Ug e }Ug v YU P c e U X e X a X Q ed V w g Ug Q } Qg XUg U P v e U g n U v U V e Ugg m e y Qg ag Q Y m Q g Ug Q } Qg XUg U P v g Q p h Q a U v XU 0ixRx4iF`"iwX(cv d~ 1k0bwai0R04bFT| T4x@"x 0ixRxFbwXdx4(q|1u0"F{hav UUgg v a eU P vg V U P n X e Y Q n vU V Q XU v X eg U V m e Ugg m e Y U P w XUgg mg Ug Q UgU P Q P aU v e X U V e iihb10(0i1wni"|"iB"1x"XW1EFiwi"FWhab1g 04bF"mwvbvTwyiFi"zxibERi@hgi1bRkihxQ s q Ugg m e y Qg U P g c ygU Y eg Q gUU V Q e Q n v v S eg X Qg v egU SU f e q y Q a QgU SU gU S e X v Q SgU V e a U Q U }U x14bFT4d|"i1eBi1i1EiiRBviw1B"wX{hab} x r1E|{Tbx1iB1BR1weiR10i0Wg e1weiR@w"ixt1wew@o|"i1ewX1wni1@w"i$f 0FWb"hg"$FbWi@0ix4g{wXzbwX({weFbxhQb1zb X v v e } v P Q X v v YU y Qg U P g c v vg e v P P v U e } } m P v P f e QggU } a Ug Q c v U P v y vg U S v a Qg U P F(hR1&Rphab1ji"wm1(1wSii1`i04F"Ft1Q xB1W1wehT4|bix(wv"xhQixhgb{weiWjx1e XU v X eg U V m e f gU g g Q U v Q Y e g Q } Y eg Q c e g e e X v v YU y Qg U P c e y v V vg Q S v a vgU } Q c e X v w ps e U Qg y X Q Xs q vU P e } y P v P Yg X eg m eg U X gUU a U P U V m e f w ps c e y vg Q v V U P U S eg } e 10u|11F`xHh0"ixW{EhbiR{Fwa"1b1wexQ 0ibbiWtha"1EEF0sz(Bxt xRX"bi1b U4b1iF0|bixbi0RiuT|weRo{FQ Tjtwy"1u"$E0Ux"wX"quxc 1"{1Wxq4hgbdf gg m e w ps U P c e Y mgg U } y Qg v Qg a e e n yQ f Xe U P q vgQ v V t r p h UgQ Q P me VQ P v P wybz1wei{"1FkhboXB1F1we dk{"vF0b} x{ix"iF"1xwg{1$1$x4(W(x4i0W|1$1T|p X t q X v v a a Q eg v P v Qg X eg f Q P v f m V XUUg c e eg Q }gU X m eg Q v } e Q g Q t U } y QggU } g e U yQ wybwVFW&ix1we iFWu1wni"i0bwXwvhg1i ~ 1"1 fc1o|1nbvFT""$4b1iT4|1{B40ba v e } Ug Q X v v e } P vg U P Q U v Y e q Ugg m e e h h P vg V U S Q P Q P Ugg m e y Qg XU e a Q g UU w ps U P X Q P w g eg c Q y V d a X Q V U q q U P d w g Q v n mg e e y Q h y v U P v U Pg 10pz(bxRi11QukR"xR1u u b"i x11x vT@ixjf 0wU1b11T&&wubwXk0b4g ~ wXvF"1jxvWb{ X iFFnbixo0bwX"1j("i "idwU"x1"1v "F|(Rwm@{01z1uX10vdi"wmg v X m eg c e gU V Y m gU P vg V Ug Q aU v V Y eg e f U P U P e V Qg Q } Y eg U v m l gU g Q eg e y w v x w p gU P v P f e q w t a X Q q w g Q n X m e y U P Xs q X v Q SgU V e P v f U f n vUUg n Q e U 4hgbdbbxF11x0 b1xTx"kBzE"1xF1Fx(0 b1x{T" x&b"11ubtHr1weii"1pi$dwjobwX1x1Fweg t eUgg m e gU P e e f } h P v f aU vg e U v m l Ug Q g Q e f U P P v P f v ed gU X v S e U kBiW04b1id"ix@j($"wmb| z~ $0ihQ{ii1Q "Fdixix4(kj($"i@hg"$@wXF411FxEb"wX1"P H U"1dbdF0z(@ d0ihQ{iiFywxhQiixW$4b1id"ixe udFwv`(bx"wX"qusqiFfedb1@`W$1TRP X e U P e w ps P v f aU vg e Q v XU e } Ugg m e U P c p h U YU y yg Q v V t r p h g c U a Q Y XUU V U SQ A 49 I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$" 
255

Y x Y p f QQ r x f T h3qs`SGs3Y 4`qG3hSWAShW3vhfyhWA%hf v R Y T d Y Y x db V w Q e V Y n Y f VQ Y e V Y p f V p f f x Y x V e f T Y x a d V x Y a fb d V f`hG4qhgss3m`W3U`h%4WAw sUS4`WuqsfG4W Y f Q R x v V Y p f b e V Y d e Y b d Y p d f sW3gW444 f 4f443 f qgG443 f Q Y d Y x d Rb f V e e Y x v R Tb R T x R e f T Y x f x e X Y p f w R e w s r e qgC`4Shffsh%Gy%4hGsh3`h1%Cqs4sqf4 p R WG s1`I'%h%WS4Sh4qsU4mWf`i3s4sWqp db e Y f x Q T n Y p f p R db b e R Y p f f R v V f V p h x e V f x f R d44Shh4As3m4%%G3cSWA4q%4s4j44sfs3h%he Y eb x x V w x x YQ d Y T d Y Y x db V w p R R e V p R Y eb f V f d Y x Y e Y Y e V d d d d d x e V f x e Rp Y p f d d Y p a Y e V e sW%3FUfqG31sW1jq4"qs%t3sWs"p nUSW3%qsuShsA4hW%qGhs3mqu%4%h4V db e V Y d Y p f db e f Y w R f R p R e f T Y x Y Y a v Q d R p T x x a Y Xb f d Y ab Y v d V T p Tb p t d Y dbQ d Rb x xb w Y R d p fb t d 3%sGUymoWoyqut4qSSiGhhyqchu`4u43u1 Y rbQ e V f x V GSW13UyfsqyWs4Shh4jiWS%WqY v a Y ab e R e xb e V f x Y eb x x V w p R YQ w V l e Y p f R d r x e V f x f R p Y p f QQ V f R d Y e V x e V f x Y dbQ n d Rb x xb w qsW%utsW1W%%hSWvW%ghWCWsgqSSq4hhy e V f x V p T x p R YQ w V l Y d V x 3Wsqs`WCS%WqWuyb 43c1`W4GW9 fF3 SSUfUkWU`hq%mU`v%hf YQQb e R T p R a f x f d Y a d Y Y a db Y p p b t f d Y w Y Y e V db d v sf uFAs4WqqqW444 f WAq4g3W%tWhT a e V w V T Y x e V Q V e f d Y Y p f db a Y f T Y f Y a Y e V x e V f x d Rb x xb w Y t Y d p R Q V f R f V Vb e Y %hoSi3h3hUsWsW1o4Ssy%AWF4WsWsASshf n s4Sh3hqj4Shffs`4hiWsu4SsSAg%hWWR b e T d Rb f T Y f Y a Y eb f V e e Y x d R T p fb h x e V f x d Rb x xb w Y e Y p f R p Y eb f V Tb a db d x Y T e R x e Y r V Y t Q V d Rb fb a a V Y e V Y e Y p f f v d x e V f GsfyU``q4hC4W3FW`4ShU%4sWs%hCqqmsW1x Y d t R d r Y p f x Y a Q T db x T Y x e V Q V e f d Y T Y p f db x Y T e R x Y rb jufqUgqhk3U``3hsW``WssG%hjkq4hg4SQ n f db R V Y p f QQ V p R Y e e x r x e V f x Y dbQ d Rb x xb w Y e Y p f 4mi4j%hSWuWoGh%sW1qSS)4SsSAv%hWR db f T Y f Y a e Rp Y T e R x Q Vb f d Y f R e Y p f R d V xb d d Rb f e R QQ V w x qSh3sqC4"s%4WysGhWmgqh4qWgymGs4mSWAhV a Y x Y t p Tb p t p R d V w d Rb x xb w Y V Y p x Y T e R x Y x Y p 3`A)syquuW4A)4SsSA4q'3s%4h3%hf p R f x V YQ f V Y e V Y e Y p WoA143SgfhWqh%kfGGW fv q%W9u1`Ua e Y d d V h f x v a Y db VQ l Y e Y f f Y v t R d Y e V V e f T Y x x x YQ Y e f V Yp a Y f d Y 3qSWy%Uosmf%)h4ih3`his3sqsf3"3sGx n Y e Y p f Y x V T Y v d Rb f V T b x x VQ T a Y a Q Y a V p p Tb p t x Y T e R hqqsC`Wmq4ShfmX hh4yiq%S4%iS%''3s%4x a Y e V ep db Q V e Y e Y x d d Rb fb a a V d 3h4s"US4s4kGsUqGI 4"GqGfqgW1h e Rp x f d R T t R d e V f x YQ db x V e Rp Q x Rb e Y e a Y f d R T p Tb p t d d YQ w V l S4%SCGS`4Sh%o3hG%4syqumuG3tWS%WqY d V x r a Y f T Y f Y a Y v d V T x e V f x Y x Y p f p R Y e R w d x Y R T x YQ Y Wqt3h3hqu`WWsu3%hWs4'3m4h3Shf Y e VQ d R x w Y f x x r a db p Y v V f V a db V wb n R e f T Y x db w R GhWy4AhUhj`%`Cfqv%GWASWsh3`hqS4T n Y p f p fb h fQ T b a xb p Tb p t U%q%hhi% qAyyquq%Sh4W3FA4fR w db f d R T r V Y t V e Y e 4G3 g3qSSq4Gh%kWc4Sh3hqgqsqG%%mUa x w r x Y dbQ a V R e v p R d Rb f T Y f Y a Y p f d R x a d Y Y d Rb f V T Xb f d Y ab Y p 4ShWmhGqSq'3q3 uv 4%3`X sGUy4'sy%'Fq4R x V p a Y b f d Y ab x V t p Tb p t Y d f x Y f db Vp Y p x e V f x Y p f d V p f sGSWCq'%WsA3c1qs4%hF qWA43fV db V w f x V YQ f v Y V e Y e V db e Y f db Vp Y e V x e V f x Y x Y p f Y x V T Y v Y f YQ w R T d gGWs4SsGSWAsWsW13%hAh%43ms%4%Sb Q db V f e Y T xb x e V f x h p R f d R T Y p f d SqSWshSCWs'WvG%4jqsfqGW f WA%W a e V w V p R d Rb x Q T d R T Y p f db f R a Y f db R a V Y eQ V x V t x V d e Y e Y t R WGh%S%4%hqSq4qsGS4mjq43hSW4u4%4Ff 4Shfy%q4mW1iq%44j%hWvsm13sG4Sh%j%hv4"p d Rb f VQ R e V f x d R Y p f p R x e Y v w Y w f x Y f p b e v Y p f e R Y f YQ w R T Q Y rbQ xb f d R T f d Y e e T Y p e Y f x Q T Q V Tb f Y p f R s%44SSy'q4uGhsqvq'%s1`WyhqsWmp xb p f p R x x V w Q Vb fb db Y p f Y b e Y f T V e V p T e Y f f Y v R f Q Y p aQ R yqsWvs4AFWyhqSj%hSss4sW`gsmisS%)yq4T x T Y x e V Q V e f d Y T Y p f db x e V f x f R p Y p f p R x x d Y T Y f YQ w R T 3W%Whg%hiSuW1gWq%hkWu``hSq4r } } } | | } | ~ } y y f4mU11tffj`1)41fF%`j`1'qA4ffFt fx

p R e p f d Rb f Q R e Y e VQQ Y f x Y p f a d V x b a V e e Y f x Q T Q Vb fb db Y p 4qGh%hGsUS44k4SShCqh%WF%SUGss1`tWyh%qsf d3s1`So%hgW4Ss`q"h4A'W`Gs4iqs3sGi%h)sf e Y f x Q T Y p f p R d Rb f T d p x x V w Q V db b e R Y p f d e Y f d Y T Y p f R Y T d V f xb a Y p f a d V x x V w Q Vb fb db Y p f R f d Rb b a a V db d f d R %4sSq)qs`Ws4AgWyh%%hsi4Ssf qqGSG%4T n4ohSq%UWsoUSqGhqjqscsS%4As4Ch3sU%4V T V R f db db r V f v w YQ v R e Y p f Q Y f YQ w R T Y e R w x x Y e a a R f a Yb e f Y e V p Y p x x V w r e V a Q V e f d Y T Y p f e V Y d e Y f x Q h3Shsi4%iq'Fh4jh4qgWhGqsjW3%h%ST e V f x V p R Y f Vp Y p f a Y e Y ab x d R T Y e x V p f e V t x Yb Y f e R WsgWghf%h3sUy`43hG%kG44WW'm3S4hh4 WySb T d VQQ v hGFSsGSW`WqsWq43Gfu r e R t Tb f Q V d V e Y p f R d V d e Y e Y t R Rb e V d Y T x xb p f f x db V V f d Y w e V d V x w Y Y x T d V Y e T xb %ShW`sS%h'1`4GW'q4sWiW'jhj%4%hsSqa xb p f d x h f x a Y e R w t Yp x Q d yqsGuA1%qCh4A"F`SqqqGW9 fk WA%Wv a e V w V d Rb xb e Y e f x VQ Y p f w R ep 4Sysv4yj%h4s3 W3W%4ssGqs)Ssfssf d T Y x e V Q V e f d Y T Y p f db a Y f V e n d Y T d R T d x e V f x w bQ Y p a Y f T Y f Y a p R e Y v w d QQ V w x Q Y eb f VQ Y q`4%Ws%S%3h3hqWm%4A4Ssfyse Y p f R f a Y e V w R T d Y p h x x V w Q V f R f a Y eb Y e Y p f p T V Y e R %hhsW`4)q%hGAWWh3hSq3hg%hi4ssf 3 g4%g`1`s1`tWyhqSgqsjS3SSs4%WFmv Y p Y v f x w e Y f x Q T Q Vb fb db Y p f db x YQ Tb f e V p R e Y n w d Y p f d Y x V T d V d x YQ w V l Y Y e V x Y T e R x Y p f qv%hU4GWqI%3%W%gsW3q4h3cvqh" Y V f x Y p f a Y p T V Y e Y e V p x Y T e R x Y x R p 4W1FqsC%43hFG%3s%4'qut c G3h R f p R Y d R e Y p f R d V a d V x YQ w V l Y Y e V w R ep d Wv%4u%hW%W%W's3SqAWqChW434s"F%G d4UF9o4W1%hisGG4cq43hSWyqus3Y r d Y V f x h Y p f R f a Y eQ R e Y a V Y eQ V p Tb p t x f T n v q4R c WuAsf c 43hhGAtWSsqSWq4`4Shfy%q4m R x Y x x V w Q Vb fb db p R Y d R d x d Rb f VQ R R t f p R Y eb f V Tb a db Y v R f w Y Y x 1k44ShfyU``hhq4W f' WA%W9Wsx a e V w V x e V f Y Y p f d R w V a Y f T Y f Y a x R e n v x R t f Y p f d f T Vp d %hiqGW3s3sq%%4s4U%1j%h4I Y V f x h Y p f Q ab V e p T V Y e R 44W1v%hjqyqW43hAhf c 43 h4jWSsqSb x x V w Q Vb fb d p R x e V f x p b t d x YQ Tb f e V Q Vb fb db p R e Y v w d Y e VQ e Y W's4sisf uCW3SyhhW%Wyh%4'mqUgGhWyjs4e V x Y x R F3Gm%qx o 4h%SCqs4hGAtWyhqSqWG4W p R e Y f x Q T Y p f p R x x V w Q Vb fb db d r fF3 b t c 43 s4vh%qjs4A'4R c 4WR w T Y e R T Y x d Y a e Y e V e p e Y f x Q T Y p f p R x x V w Q Vb fb db d V p fb t d Y w Y e f l Y Y e V x d Rb fb a d R h%SqsUWh4AUWyhqSWgs'3q3AshUhW`4ShU`4T a Y eb Y e Y p f f V p f f v d e Y f x Q T T Y x e V Q V e f d Y T Y p f p R db b e 3hSqsqsf`hqqGs1`u3W%`WssG%htWA4Ss4R Y p f e Rp x Y f V ab a d V T x V a Y a e V Y e Y v d V T x d Rb f VQ wb x Y w R %hgGgsf%Sq%WgG3UWGhq`W%GsfyqSq4hx f V p f d Rb x Q T d R T Y p f R f Y w V T Y p f%s4S`S%4qsgsAWq'4 s s4UW3%kh1`SWsx r e e V Y d e Y f x Q T e V f V d R d Rb f Tb ep Q V Tb w V d a Y p f p R d T R f bb a V e fb v e G4sSs"WyW`UqsW3m4ih q9uSUGsvqs4R QWyh%uGhtU%WF c 3i`WV c 433s4A'Ght qa Vb fb db f d Y e Y b a a d V a d x Y x x V w f d Y e Y b e Rp d x d Rb f VQ wb x Q V e Y e Y x Y a V w Y e V p xb e e 4"m`4ShWS%SW4q4Ag4`fG4WgSshGR Sb w f d Y w Y f V f x xb p f f x Y f R f e Y a e R d Y wb f Y a db T Y x e V Q V e f d Y GhWs1'yqsF13shFqsGjIU4ShqqSu3W%mWhGT Y p f p T V Y e d V T T d V p f e Y x RQ T YQ f fbQ V a Y w e Rp a d V e Y f x Q %hg43hjWtv4W%skgsSvA3s4"`WFh%ST x Y p T e r Y p f d V p f Y eb x x V w Y e R w Q f d V T Xb d b x x e Y f x Q T f V p 3qs%h)W`h4hhGAAs4oShG4mq4Shsh%Sgf%sf x Y a Q T d R T Y Y Q v e V YQ T d Y p f p R aQ Y X Q V ab f d R e f x Y p f d 3U``4W4S%qFW3`qhWAS`%4qyhC%4hs1qsASb d Rb f V e R V e Y Y p f Y eb e e x R f 4Shf4mWfqhg4SUh%jsF o 43uu4q4%C4hhGAw p R d Y Y eb x x V Y v f x w x w e Rp e Y f x Q T Y p f p Tb p t db a RQ T f d Y e V Y p f p mA%gjh4"s1`iqsS%'S%4SAGsW%q%hWR x x V w Q Vb fb db Y p f f V p f xb d Rb fb a d R T db V w Y p x e V f x Y eb x x V hGAm4SsqSgqh'f`hky4ShU`4SWAg%UWsg4hhGAw f x R w x b p R Y wb f YpbQ Y p f db p fb t e Y f d Y T Y p f R f db Q V eb x R 1GAjf CWoSh"Si%hi%hujhoqs)hW%sf T f V a Y w e Rp e Y f x Q T V e Rp x d Rb fb a d R T Y p f p R d Rb xb e Y e tA4Cfi3s4"s1`jC4"u`4sU`4qsWq4SyhV v f d Y w e V xb p f a Y a e V T xb a a e V p e Y w T GqGhWyqs3UWhyq'qGW9)s4%sgtWAmhU3 fb x d Y a Q V d Rb f Y T l Y d V Y e V p aQ R p x w Q T Q Vb b db Y p f d d Y p 1`q%4%4ShqqAWu4%AS%4%jAqS`Wysf %%h4qsf a d V d e Y f d Y T Y p f w R ep T d V p f Y e R w a Y w e Rp Y e V p f R %WsGqs4h"jmSqW`hjh4uAh4"AG%W%d aQ R T e Y f x Q yq4)h%ST s s4Uqscf%sSWyqugW4GW9 fq r e Y p f f V p f wb VQ T R p t d e Y b v db V V a YbQ V xb f d Y w e V Y w V x Y p Y T d V f xb 4SSWW3S%q4jyAG%4h4q4hoq'4`WsSqa Y e VQ R R f f V a Y w e Rp Y p f pb T Y x e V Q V e f d Y T Y p f db p fb t a e V 4sWysfj3s4"%h'3hsW``4shC%h%h'AsW't n db w Y p f f e R x d V e f R f x e V f x d R Y eb x x V w p R Y wb f YpbQ Y p qSqs4m`WhAsWsj%q4G4Ss4AWAsSqsf d V p f e Y d RQ Y r V f aQ R t b f V p f a Y e V x V p xb e e W%sGq4Sg44siyq4vf Cf`hq3q4sWC4`WG431gyhs4R } } ~ } ~ } | { y zy %u4`fWWyfuGIWjqA%"%qyWx

4G4W9 fF3 b 44ShWS%Sx w p R d Rb f VQ wb Y p f db x V x d Rb fb a d R T x x V w Y w Y e f l Y db eb Y e f R d d YQ vb x VQ %hj4F`4ShU`4s4AhsqSs%G3hFWq4SqS`Wyq Y e R w aQ Y X Q V ab f Tb f T VQ V Y p f db e Y f x Q T f T V w R T V p R d Rb f s4S`%4qyh'yh4yWGgqsjSs1`4`4W4ShUSQ n R x xb a Y p f v e Y f x Q T e V f x Y eb x x V w d Q V e f d Y T Y p f p R db b e R Y p WsyUqsks1`4s14Ss4A49WssG%htWS4ShGu%hf x Y r V w r e R t e Y f f VQ xb p 34WAcs4Fsfyyq'444W fg w b Wu4ShWS%Sx p R d Rb f VQ wb Y p f db x V x d Rb fb a d R T x x V w Y w Y e f l Y db eb Y e f R d d YQ vb x VQ %hj4F`4ShU`4s4AhsqSs%G3hFWq4SqS`Wyq Y e R w aQ Y X Q V ab f Tb f T VQ V Y p f db e Y f x Q T f T V w R T V p R d Rb f s4S`%4qyh'yh4yWGgqsjSs1`4`4W4ShUSQ n R x xb a Y p f v e Y f x Q T e V f x Y eb x x V w d Q V e f d Y T Y p f p R db b e R Y p WsyUqsks1`4s14Ss4A49WssG%htWS4ShGu%hf x Y r V w r e R t e Y f f VQ xb p 34WAcs4Fsfyyq' s sGgW1qSy%h43hWqd r e p R fb db Tb e Y p f p T V Y e f R QSS'm3%W1yUvSqs`WghGAC%h4h"m3%qsGSqj%hf Qb t d x Y T d V f xb a eb Y p f a d V x x V w eb Y p f w R ep d f YQ f db Y p a d V x Y p T e r Y p f d x e Y f x Q T e V f x d t R d r Y p f d Rb x Q T d R T e V a d `W3qsuqh3s1`SkWsg'f%'qs4S%S`4CsW%%4R nhFU 3cSqsf umkW%W1yqmW`%CsGsjWT T Y x V x r e db p b t T p R Y T d V f xb a Q V d X V p T V Y e d V T tj q WSU4h4AIyW`jsf utsGqs4s"tjW p R bb a V e x x V w npQ V p V p b t d e Y f d Y T Y p f w R ep T db p fb Sqs't c 3u3s4Ajh'Ah4ms1`SFWsf`huU`T x Y x x V w p fb t d e R v x e Y f x Q T e V f x f V p f Y a Q n d R T Y p e Y f x Q T Y p f p R Y wb f Q V eb x db Y p f db e a x x RQ x x V U4q'3s1`S'qsqWSh44sSqh%Sqh%sqqhs4Aw

w s r e fsGv a Y f e xb a x V t p Tb p t e Y f x Q T V p T x p R Y e R T db db V w Y e Y p 3hqqsyU4'qyquuh%Sjo%hkWvhGqSqSWjsqsf xb d s r e R f Y x RQ T e Y e d x e V f x Y eb x x V w p R d Rb f V e 43uCStG4ihhGSqsGmWsg4Ss4A4oGsfssf nq`4mW%4SsUqksjhi4U`h1FGhsqqsg3hffSh4w d Y T d R T Q V d Rb f Y T l Y x fb p fb h a f x f d Y e e T Y p f a Y f V eb f R p Tb p t d YQ b fQ w a Y e V Y V Q d f R d e R YQ db x x V sy%'G4%ShqsW3`%W)43Sq3Wq4'S4%S4't Y T e R x d Rb x xb w Y Y p f e Y p f Y p t r T Y p T R f e Y a e R db aQ Y X Q V T R s%4hGhhy%hv%hquis3qihgU4)ScS`WUSQ

P D 0 ( 6  E E D B @ 8 7 6 5  2 0 ( & $ !      30 1G2 3I1H GFCA'9'1431)'%# "

256

ANNEXE C. ARTICLE THE NATURE OF IRS 13

9 9 3 8 I ' #  SeS 4te"710w8 0rt3 r1 o"e trwFgHF$ 3 6 3  D2% ) pfs XafXs c csd Rfa f id U R Ud v w s cdX dw i f id y w X vX Rf tgt1"VettFf"gY`RY1`wtp`c"VFSq1teVfVtS1"g"a 3C' # 3  3 ! 3 DC ' %23% 1 et(wk 0r2 IVfre&FYh44H({ `VvVtgeSz`X"gx`R&0gett(Sv``y1zF1`e4V1@t1q y f id esf c U id U Tfd csdX U X c X c e csf Rf W fs w C I3 # 3 D 9  23 ' 3 D sf idf W wd T R v w u c Ud c f U fs c Xfas v wX c v U U T U f i (ptkFq@I "eSq q" I egtwHro"4% HYS o 0tE! "g11t""1gq``ebagp@"Rrg1B4V1gheSrbT`rS`R@Vtd A # P  gre!(V1 0e0VkrVg(tE! 9 D 3 `epSgt"V1zgY`VS`fqg"`vf"ggeq4SqtF`Recy c dc id s Q R wa wd f i U sfdX a f id yw Xsf u f sfd U P  ' er1 pr0r"t% (( r8 r""Y1 1!SI egp q q" I egtw  4FEe r1b4E! "gekS1`wgeSFghte111Sp$rptg7 d "`RVgu 5 I3 2   23 3 D ! 8 3 f id y w X R Ud w sf ws s wsdf T U if i $h"eF1e"phrw1Vq d k rj F`UBg1g0`ftd m u T R c s f W Rf ~ d f f waX f 9 9 @p"0eFr8 EG 6 p c R w fdX X ~ u R c T U if u Xsfdf wsda f X p l T fs c r y 9  3 ) D I3  CC p!wF4tFF4H% E@ E4HE@ 4H0D t4F4E! pk1t(VxeS`RVgtg1tg0g(pz0ge4sk(`Ri dU C2  6 3 I3 ps cf R Tff R U X U T vdX fXf i Xdaf u w m l T R c l m l ge0V0"S``$E0fV"t(k0"r4r110hSey1r j AP9 #  9 I gp1 eSq {"z($! 44t k4trw VF$ C0(Hqw wt4F4E! "g$eVtg0Sgg0`fg"gp0f&0g0gr"ggSrFxbT"1VgX I'22 3 ' 3 3CC D 6 3 I3 f id y w vsdaf X XX fs vd c y Tfdaf f f id s Q R wa v w i XTRcu TR VSe""ec "gR1ggt40g1`gdg"f j {0g`sVts f id U waX wsdaf X f U UX R Tf U v nf # A A P  !re0Ve& e0VSt% (HYV B orV4 2 ' ' 5 V C F4SV kS0HF4rtSS q 1&SI eS" 6 1E! 3 I'% I3 3   D bXx`TS1"@1t1qtS1wh r"1xrr1z11`wgV`v1w U f Q da c wa sf c ia vX s y X c T w w W X c R Ud  4(gV E r8 ""Y q qB" I e4wA w kq4t0G t 0gh1`c"1"fFt0g`s"F0g`Xf1"et{(z"g`Rzr1gFtg0SgX C3 ! 5 I3 2  23 3 pXfs s v W R u Tf U v nfs U f W R cs p f id U waX wsdaf P PP  9 5 I3 2 0p001 eVV8 r""Y e m l j i U i U Xfas v wX c v U U T U f id c y w f d csdaf f1r k(d `RVtd$0t"1gSeSrbT`V"`Rf"g"`et1fgre4g0SgX f6 t "2 g qS1pte{ ot% (HYt r8 rg fg w kg4t0G t "g1bXV`e"e1pgeVtg`fVF"1gSr1 D 6  I 3 8 ' 8 3 f id R w UX c R c f w u c f id fd wa T R c s Q R wa w A P##  P9 I tVpr1 prSq {"z($! U 44tt 1 )ty1g f""FzbES fSb44Ftt 4H0 pgFg`X"gzettr&`V`uggFg@4V1grtttgbXVo1VwfR tdd I ' 2 2 % 8 6 I I ' 6 3 I 3 ! C 2 U X w U i d d c T f d a f d f T U X X w f a s v w X c s p f d f s a U T it PAA #9 r"p@rr8 e00V"b00ty{9 ! I GG3  U U d w ws U s a U c f id U X c ) # 3 % # 2 2 I ' 3 I 3  y 1w ` X Fc Wts w d f w i i YXVc t$ feX twV X o UFw & wV d U X @f fgsy "w g UYa4s4stdt p03 ((1we 10H0' e pFI t40w b tFI 14w q(C 0' etRX 1@11VVg10`T S11gs@`X""g`Rf f&V"1`RRWW tg`Rs eVge1`w`Rg1eXd f U e f id c y T id R d c f P A A '  3 I C % C  6 3 ' "e0re%rr0V 14tFfV 1 or"HC w gr0w# p U f i l s ds c X U Wfds w R U a u TfX w ws V`RBV$V111egegf`f1gg1ebc`qqgFg1g" 7 7  6 4 1 # A 88t&53 2r pr0re0rerfY I # ' I ' Rff u X c i X c e y wesfdX a s cdX f UXX c e f U U Us c y gw1Sw11$tF01rt(Svqe4($1`t1@p11`td @tVktew tFV"0F(e(' gpty2 (t $gtt4Hrz00{(pt 3 D ' D% 3C I3 # C2 ! %23% A C I3 # 3 D I tp00tkS"r0t q)"0(1t{ )0r0tgg S4tFt3I w# FeSgy"g$1o0gFg1g"$bXfFVFg`srfbXVth`RyVbnV"$bXU A I )' 6  I' I3 !  3 '2 C3 d R c R fs f id X c TfX w ws U d Rf R w U Rf U id U f v U R v # 9 A ' 9  3 ' 2  I 3 (Se!rp%0e0rw&Vk04Stert i U i daf u w U i | dX v T W U T a U | wd ~ w tba"$$40r1E`X"VX i @("VV`RV"`v"`Rre i YgBB1tsy w$41rG q 1%((p0Yq q$"0(1t kw"0w kw4tFt3I w# 1"et$tgES1`wg1pB"gw10g(te4g0SgzV$"t0gp4gX  C ) 6  C I'  I' I3 ! I )' 6 C3 f W R cs Xsf u f X v Us c f id y w Xf d csdaf X f i Tfd csd # #  p1 e" 1pg$xr BSt( G 4eV f1) w 9 ' 3 I3 3 8 6 A 9 ##  P9 I3 3 8 ' 3 6 p R w f T U m l j i y w R U UX w f id d us wd fX a eXs cd 0011 prVB"4( G 4e { Fp4t f1) w V1VEbX&10 (d e1`wgdSVthpc v t1W j gfw4etX P  9    I' I3 ! 3  )3 f UXX c ed w i y w sfdX a s cdX da c wa c y w fa RfXfs f i 01p {"p!1t41q"0(e4e 0p z4e 14C ttw 1`t1@E0eVyet(S`voeto1S1xzeqS0tVV$ 99 e0eP `0tbVtk 1 e00{B tpg3 HF4g{ 6 kt4tr2 S k V{eF1{  A % I ' ! 3 6 8 ' g k0( yr q"0H% t0G t g 1p71 bwF ! 3 !  I ' 3 3   C ' )# l s ds c X U W fds w R U a 411117Bgegf0f1t1qebc`` e S(F gt% (HV kreFS B wtFr C I C ) ' 3  ' I G u TfX w ws X c sfd Rf Uda c f id Tsc wd W U UX sfdX 9 9 9 A 3 )C 9 6 %3 I 8 I pgee0(1$eS fF4(t11 k"4 3 0S1tVh1hgEbat4F`c1Bk"gy4eptBV`R"Rgg"`va dX W w y f R s idw f e s c &ct y w $V `R`U T f s T @" ec`V R Vi U d e f y 1XdgX f e X f ew fhT1f hsgf u S v f w t 1pVrHew etV1ek1p&r4Verfqvx1`wtgYF1wWsc g4Fk4(`fdc B"Vg01t1bTu V1Vd s c v d sWf s d i X w i eX U Ua f U T cs UXX w Rf i T Rc X R Udw sf ws X 440`fgdr`wFSebcV1tk`fV`utFgo"$"e$"1`wteEg1g"B0f"v I C ) 'C 3 A9 8 65 3 DC ' %32 ) '% 3 D p Uds c f id y w X U U a f T fs vX c f f i XX c cd wd X U y "0' ( 2 FFh' 11" 47fF(t44(10H$F(C baggeSo"gfez0ftd Vw1&0tVg10o"t1@etetztdew ' % I C' 3 ' 3 DC % ) I3% 'C 3C 322 ' D CC % ' D R Ud c UdXf R c T R c X Uds c y w sf u v R cd wd X U esfdX kF0' (ygV1yF(F4H@o(Ft4thkHBz0F 1`wgpt(0x1Se0`fbateSBegVrfe4egtdg"`va 'C 3 ) ' 3 % I DC 23 8 3 DC 3 %3 )% % ) f UXX c U id y w f id f Usf T wd UXX w f c v w (E1zH% (he(S 4tVeV4gVq% F(I YH% h4FH(% h0 ' 1`t1@hbXVtye d Vt1`grg`fV`utFg1e@`T"1a 3 DC 3 ' 'C I %% )2 'C 32 3 I % I' H0$Fr(1y B((F4H% Fy(4 S1qk0$@ I g1k@egy"gt(Sv1tfe"7e4gy"gFwF1h1wtd s w f cX f id e s fdX a s cdX d w i e csd R f a f id s y d R f f R U ' D% # ! 3 DC 3 ! ' 33C '2 G 0FHe' r0q$F0 3 4E0gpY3 f4(HC yy4yFI t FI X c w R Ud v X id U s dRf G ' 3 D C ' C % I D I 3 D C C % C ' 3 C p cgXgs `v wyhs "c i d d  U 1 U`wU X pc y tsxf1X `wgXVs d X X f 1c ggfFcg w f U d w c baU 0T f`RU FR 0yF(@$e{Vrr&zF(I @FB42 (F0zr (FF0k% FD e@"1z1tgx`R`f&V1`RRW ggqe14e4("ht1o"@fV1`RyW p1wR `w`g1p"wrSgeca d Wf W c UXX % ) 'C 3 'C 3C 322 ' D 'C3% I' )3 FYB3 F 0rzY(144tFqH0(tHEq014Y s f d a f X X X a v ucsfw s C 3 # I I I D C ' C C I A 9 8 6 5 I ' p i R dU a Se d d qes 1  gd 4 c1d y X y 1c V 1yFwff4d"a R f c rT 1R e tl 1pR R gcyX gF w S tFt3I wqF@Fr( y(qY(pqB3 pq7q0z3  Vg1fB4hqqhy@els t11`f tsUw t@@`Rqe401VSV"110eSe@gQ w w W UwW X uw c id fs C I C '%3 D D 3 DC ' I C 3 3% 2 3 DC 3C ) C F0' 1) 0(t(FG F(0' I yFtYH0' t(4F (% D D C23' 3 DC ' CC% 3 DC C 3 ' # C3 ' 6 D % I X f T a U l s f w cd c T u ~ ~ u 2 Fhog4H(1FHF(otroHtr0(C h@F0' H% 0rSvSRYe11eh7ezrtd Fgd j tp" d $p j kpk w y R Uda R y XX c f i X f ias u f id s y R f cds f T R % )2 3 D 3 G ' I 2 3CG F4H% 1V`)F 4F0F43 hFe Fb)g(gE3 f4FgygpH rh 1gsqq7Fwt""vgF@V$r0V4tBVtE1w11tgV""v T cf c Rff u X c i s w y w f d U i Tfd wa f u wd X Tff r10tsegw1Skt1ESe$Sr(bXV$pt0tf"1gg"0"R Xds c sfd R v wa c Ud w T fX w ws f u T fd fdd c dXs Q eXa fXs c csd Rfa f id tge"tgF"1e`atr1o0Fg1tV F10Vp01`fU i r@0gVggptSr0gteSVe4gkVt`RU w R f c i i U i s s sp l ~ T R c s Xs cdX W R v w f id c y w f s vX fd wa u Tfdafdf T Rff V&1p"tba"$F1F111(0 V1ere"zESee11es te4("V11Vth`ef1q1tVgt`f"1qt0g0gVgu p l ~ e sp ~ f l f c i Xdaf u w XX c sf T R c fd U Tf sfd U yw X c f T R r "1erpSzBkEtSV11ep "Fpre"zBkE1``U{e11es 1rSwt0r1wgF@wEp`w"1ge`cV0gtF`RSe`f"eVzSec w R c W c a fXs c csd R fa f id U cs T R c i u T fda fdf Xs cdX f f UXX c dX w f id R w e R Uda Ud f i U i s f prheVee704eSre4gFVgR4rSe"E@ttrT t4etV`v`UViB1`t1@qFfVty"1p1`wgSRtY1`WVyt1 X fas v wX csp f ps cdX sf id w f u dX v e W U W c U R Ud wX f f id i U e s y c f id y w XafXs c csd Rfa f id U c T w 0t"1orty1``Ue4(kVteg"S""`R1e@`1`wgr`v1g0gs "gtd fr0eckw j 1VpbceBf"g"ewgt1"F1tgfVt`RyrVrtd Ud c X i U i e T fs c y Ups c f R e f f T s y Xdf Ws cd Us Us cd c T f s fX u w e Xs cdX f dp ~ w 1`ctp"ghFW"1{0ge4sr`Rge0Vr{rf1wh41ge4(dgFwthVV4pVT 01gg"1tte4(@Sr(7td i }1tst{FwtpbcV"1SBe4(ze"i | w y e R Ud v w s cdX d w csp u d i W v ws u f id U X vX Rfa R Ud v w s fdX f i f Ud Rf f id y w s w U Rf W ws f id X c i U i eXf ias u f id R c i rtHfrEFFV1tVxzyVg`RESSh1`wgpbcV"1gEebc`g"gd tstF"g$ez1td "11g""gyFx4`a"$w0V44fhVt&e"td y w c f Rfs f id y w R Ud csdX U R c w c U T vdX U i eye"txVtYe1`wgp4g"`v``ebXebXV"t(bXV$ f UXX c fs w e v us y w a s U i U e w Wc Xscf X R U 1`t1@g1@e1t1W j 1te`R"gd$r"1e4e01S1w`U``k0 R ~ j h1w j E d r1geh0g1wr0t(S`ve4(B1`t1@`f`Ug1heqet1B`f"ugX d v w u c Tf s y esfdX a s cdX f UXX c s s cf R c y w fs wa U i U UXX w R w cd c T y w f d f id W U Ws Rf e w pX w f id U R U UX w wa f v U R v X U T R c R Ud ca f v U R v X td f`fV`utFg`V1tp"Ye&g("g"RFgebc"$h{gd FgVgbX1`wtd gFg1BVFbn"V$tdSeoFwtpr`wVFbn"V$tdU R w T Ud cs ifs i U i i U X fdX X ~ u w y cd c T Usdf fpry`wgp4""gtd j FW"@td fh@t(V1tstp"fbatg e0g"`v&10 k(d VthpSgFg1tVfEe1`wS`vS1R d uesfdX a m l j i f id dc id fX w ws f e R UX a R wa pwdw i i U Tfd Ua wXX c f u wd Tff R f i f i v w iX l ege"Vgd$y0tp`cVt1$ggz0Vy"1{`VzbTVFVgwe11es R c id fs w U U U i U Xs cdX wd XfXX c s w 10&eSgg1@``e`ctdV`Rqgd$te4(gEggF@ebc1gX prH0gteSf1tg"gxF d $1 j ktd$ e4FgFgs$te4(@"VF11@rS`vS`Rgt1"e4gFVtd dd ~ afXs c csd Rfa f id R w u ~ ~ u i U csf fX w y Xs cdX W R v w R c f T a U afXs c csd Rfa f i Tf Ucd uw Tcf c cdc T ff T f Tfd wa f u wd X Uc 0V`Re4r1hrFt`seo4p"rt0tf"1gotk"`Re@ `Rte4(&"V1Fex1`wgVV`ugt(bXrq"gpSghgebaV"`RYF"RV"SQ U Xs cdX W R v w y w R Ud v Usd U T f id d c id fd c U T U X W U T R P 9 I C '%3 D D C A9 8 65 32 ) '% # !    0' (F) 0H4(G FEB@$7$4(10(&$" 

257

r R B R y u R r r r q T T @ R  5 P e @ c T T @ R x ) e 5 E 22XS2f3fFXS22U UfGfP S3UR vu XfUv`fU B R R y u R T T k R ' ' ) # c T T y R ) E 2US2v33Ub$S2z f6I2FXR oy 3USUU R r w R y u R q r r q T u R  5 D ) % c T T % R H P T v2XS2X3fbXiv2UvUS`0UR b% bXSfzUR vu R 5 P 4 ' 1  T T R H P k T R   k T % R T R E 5  n 5 ' 2G2) `If033R pi oG3GiR f5 3iR bG&63R S$632X`IG1 v6 R R 5  1 D s R q r r q T T % R H P c T T R 5 E D D ) # T T R H P i3f0G0Q$iv2$b% bz3G$R o} Y$9 6G00F$R pi o$3G{k  s q e q B R s u R q r r q R n T x ) s R P ) h 00`Up CvfvFXiv2US9fUQGSUtP  5 E R ) E ' e  ) k  5 D P 5 % D 4 # R ' 4 5 '  r r B  D P 5 % 5 E `GX$2G0S0{Fi`W4 0` &``6IUr220i`pfj P zS22lY Y$P S6R `' 00 v6GQzR zGflj R B r r q T T R e 4 5 a c T R   ) R T u R u 5 E # T u R  5 P P r 22 R w R y u R q r r q T R E 4 x ) k c T R  5 P  5 k T R  ' ' 5 i22v3fFiiv2ivz0D SGP `u vR v6`G0`{2R xi6`5 0I`F8 R US`0UR Y f$P 36GR X`' ItU G26GFfR vGGflj T u R  5 D ) % T T R e 4 5 a T R   ) R T u R u 5 E # T u R  5 P P r q R R y u R T T R 5 e 8 R  y T T i R E D  % 4 2XSif3fb$S22z3y Uv65 Ff6fk pGiiI R  z2IGQR x 2I2) R oy o2v`5 bR ($3 Gplj T T R P ) ' P y T T R 5 ' % T T # R 5 e 8 T R P t V R R  5 H P 5 ' ' 5 5 E j R B r r q 222f2G6II6tQGSS22ST SrS6&vR v$66G`ziR R V  y c T 8 R P 5 u P 5 T R  5 1 E P 5 T T i R  5 R T R 5 t 5  # T k R  5 P  P  a T u R 5 ) D GUG`0' R p6`H lph Yf`X`0F2R 66f3fR 20D 20i@ B R T  5 @ T P ) % @ u R 5 e 8 T T 2$SzR 3X9323&G2v`5 bU3y x R 5 T R E 4 ' 1 i T T k R  5 D ) % T u R 5 e T y T  5 R # 5 E S 6tfG0GQ3y S224 Uv6X6vSQGD ) ' 4 5 '   D P 5 % 5 E j R T R E 5  n 5 ' p6`00lIi6tGP S2l$632X`IG1 R 3y x c T T R2v`5 bR o2265 bR YS 6tYR k oii00YtYR v8 UYvi) 20i@ 5 e 8 T # R 5 e 8 T R 5 T T i R E D  % 4 T T @ R x ) D v R r B R y u R r T T 8 R P ) i c T j R  5 1 # T T u R 5  ) V H P ) S22fSUb$S2$7 Gz23 &XR QU6`G&UR 3y vf20W2G2z R oi 36I`G0&R 3y 6I`IGhR l $6GR xzWV S@ T T y R D D 5 P  # T T k R D ' 5   ) T T k R P 5 E % T R P ) r B r B r E n )  D '  P R  s q vv2XG0IIfU6I0`Urp e q B R s u R q r r q R n T x ) s R P ) h 5 E R ) E ' e  ) k  5 D P 5 vzobliv2lS9XUFzSf2&xP  6G$$2G0S0{Q`0i6% 4 0` &3`6`00f22fIi6FfrP 3S2Svz`G$0G1 D 4 # R ' 4 5 '  r r B  D P 5 % 5 E j R B r r q T R E 5  n 5 ' c T R  5 E ) D  E % T R 5 e 8 T T @ R x ) D @ T T s R x ) 4 bR xb6`GG2' 3QR x2265 bYR 8 Ubvi) 20iR l o25 S6S@ q B R T  5 @ T P ) % @ vXSzR 3X9323&u R2v`5 b3y S 6tfGIG&3y S0224 U 5 e 8 T T c R 5 T R E 4 ' 1 i T T k R  5 D ) % T u R 5 e T y T  5 R # 5 E D ) ' 4 5 '   D P 5 % 5 E j R T T R 5 e oiXiiS &QG$6`00XIi6&GP 3S22U3y 32v`5 b8 q w R B B R y u R T T u R 1 c T T R 5 e 8 R T # R 5 e 22$S2U3fFUS22zY Ugfl$R 3y Uv65 FGvy oU2v`5 b8

q w R w vz3 R y u R q r r q T T T k R 5 c T i R ) ) T T @ R D  ' H 5 D  ) 3fFiivvo# vy $P bz(2R oi5 2ilvR Uv60Yf65 260Il R r r R 7 @ R E 4 t P 5  1 T u c P 5 T T 2XS2vfY3U$f3`00 fFXhCU`zU3y T  5 R V ) P )  D ' u D |  D 5 V )  5  5 D P7 R r r r q R '  ) E Xiv2G2IFb64 03Q(P I0W20`gIi9P S2`0G0D n 1 ) 4 c R P P V ) y T u R E E 8 T R D 5 T T 8 R x )  D 5 G26$GrgFR fG4 G4 FR 6Xt AR o# U00l  s q e q B R s u R q r r q R n T x ) s R P ) IUrp WvGob$ivXS93UFUSf2&h P `f$2fIS0{ti`4 0QQ`6`00f220i`%  5 E R ) E ' e  ) k  5 D P 5 % D 4 # R ' 4 5 '  r r B  D P 5 5 E j GP o$2pXoii) ipR l SWX CR FoXrG R t B r r q T @ R x ) D @ c T T y R   T j R   V 1 g   e `IIfCd u c u R B r r q T R   ) c T T y R   T j R   V 1 vf6&QS226' ItR l 3X R FXrG R B R u c 22 S22Gv&Fu R r r q T R D 1 8 c T R   ) T T y R   T j R   V 1 2S23vH FR x`' 0ItR l 3$ AR QXrG R R  5 H P 5 ' ' 5 5 E j R B r r q R '  ) E D 1 ) 4 c T u R H  5 t P  5 D 222l2G6I0`tpGY$S22z`0G3v`WXR X30`US0`0i@ R FW6`3fG`5 FrR r60Gf`zWR p$frgFrR pC$f2b8 T j R  5  P u T R D  5 P E 5 T 8 R P P V ) y T i R P 1 ) R &2Up2`FR &60QR Qf`6G6FR vbf63vz3R FQ6IF| T R 5 t V ) 4 T u R D  e 4 T 8 R 5 u P 5 # T 8 R 5  ) E 4 @ T j R D D  s q e q B R s u R q r r q R n T x ) s R P ) IUrp WvGob$ivXS93UFUSf2&h P `f$2fIS0{ti`4 0QQ`6`00f220i`%  5 E R ) E ' e  ) k  5 D P 5 % D 4 # R ' 4 5 '  r r B  D P 5 5 E j R B r r q T 8 R 5  ) E 4 @ c T u R D  e 4 T 8 R 5 u P 5 # T j R D D GP S22f`G2SR v6I2R g6`f`QR F6IF| q R B q R y u R T 8 R 5 u P 5 # c T u R D  e 4 T j R D D vXS2f3Ub$S2$Ug6`f`QUR vfI2XR Ff6IF| r r w R q B R u c u R w T T 2zi2v3F$iv2$l 8 R e u   1 h c T i R u D 1 T 8 R 5 u P 5 # T u R 5 t t  h T R )   w fS`0D 03f&R p3zR vf`6G6FR 2UfQR 300I6&s B w R q R u c u R T T j R t 5 # c T @ R u D  i T i R u D 1 T iv2FS3238 Fi25 f6F2R USUF2R p0G$iR 8 R 5 u P 5 # T u R 5 t t  h T T 8 R e u   1 h T T i R  f`6G6FXR vf2UGQf GgS60D 0SU&UR 3y oU6`5 QGR X300I6&s y T R )   w q q R R y u R B r r q T h T R ' P ) j T T # R  5 e 22XS2f3Ub$S2XpGUi5 If6UR v8 oUv4 b8 R 8 xz`' ItzR l oX0 W0&$R oi Y$$2I' `2H `ffR U@ z`2i0a T T R   ) T T # R 5 V  # T T R P ) 5 5 T T R ' ) )  R vs QSfoR k o3oR Ch of2oR o3Gft T T k R D  P  a T T R u D 1 a T T R ) P H a T T R ) P 1 R w 2$i2v R y u R w T 8 R 5 u P 5 # c T u R D  e 4 T T R  5 8 T T h R P 5 D P 5 3fFiivSv2f`6G`Q2R 2I2iR 3y xvX`bvR x C$60i6t r B R r r R 7 S$iv2vGY3@ T 4 )  R E 4 t P 5  1 T u c P 5 T T T  5 R V ) P )  D ' u D 2i0i$fS60I GQSvf6z33y i$6vio2G20IIbU4 3F| P 0IW2gI`0i9P $S22z3$6`Ut `33$R CX6I`v4 zR b%  D 5 V )  5  5 D P7 R r r r q T y R  5 5  E 4 @ c T h R D  5 e a T R  E 5 ) 8 T T s R 5 D D E j T R u 4 5 j T R  5 1 E P 5 T @ R 5 H P 5 25 0G`iFfR u vUIf$R $i6`$R X6ffG6I' XR Uzg62G6t R T  5 @ T P ) % @ 2zivR 6`SX932(3Qu R P  ) E D yn  P a T y c 5 E H 1 5  a P x T T  5 R ' P ' P 5 i $f0Y6QSU $36G2G60$Y$XU6G2) 0G`WV QrB P CIvx GFCGGP 2rxP lS2R 3X l6' 00t 5 '  5 P 5 E D H H V 7 R r r r q T T y R   R T R   ) R r R y u R B T R   ) 2vUSvf3fFUS22zxX6' 00t B r q R R y u R r r q T T y R )   c T y R )  E P  5 % T u R D 5 P ) 2lvSv23fbi2S2v8 3v33(vR 3v30f4 f0`vR vvf0f2t q q r r B r E 2v222UG n )  D ' R B r r q T T @ R D  ' H 5 D  ) IS22 U6IYU`5 2`00R k U$ tt c T T T @ R P 4 R vliv6 R 63S932xGQ2$G0Y6Q3f i3G6G2G60CviX65 T  5 @ T P ) % @ u R P  ) E D yn  P a T y c 5 E H 1 5  a P x T T  R`G2) IG6WV QvtP FI0`vx fFbf0vGP 2r9rP GS22GR 3f$ ' P ' P 5 i B 5 '  5 P 5 E D H H V7 R r r r q T T y R   q w R R y u R T @ R E ) @ c T R t  1 ) % T R H 2iXSvG3fb$S22XUUvS3UU$P G0G2XR X$P v q R R y u R B T 8 R 5 u P 5 # c T u R 5 t t  h T i R u D 1 $S3G3fFXS22Xvff6`G6QUR vf2UG&3R pG203z q r q R B R y u R vXS2vf3fb$iv26 Tvs xbiiSIF&R 3y 60IGQR 3FGf QR XQ20f2Xt z T T R t E ' u c T T k R D ' 5   ) T T y R  5 E T @ R 5 D P ) R w q R 7 @ R H  5 t 5   v$2ivXY3X30`U63u T vtXt&00Wf6GS`X2`I65 `x4 3F| T u T  5 R )   ) V ) j  P  ) j ) ' 5 ) 4 ' 5 j D P iv2v! R oQ$f2FFR Q3vH FQR l 3&$`QR ovg R w    T i R P 1 ) 8 T R D 1 8 T T y R P  ~5 } T | R R w R y u R T T 8 R e u   1 v$vi2G3fb$iv2Xl ffS`D 03f&h c T u R H  5 t P  5 D @ T j R t 5 # T @ R u D  i T R 5 u 5 k T XR U30`US0S$R F$5 f`Q$R UX20fQUR Xf`60D `{GR vs R20IfR Y z$r0&n G26`l$R o$6v20&$R x$20W% 5 D D E j T T R P V  # P ) 4 4 j T y R  5 e )  h T R P )  5 V R oF3$R 3IIQQR vQI2&R vQf6`G6Q&R &2UG&h T i R u D 1 T R )   w s T u R D  e 4 T 8 R 5 u P 5 # T u R 5 t t   s q e q B R s u R q r r IXrp WvGob$iv2q RS9UFSf2P 6G$2G0S0ml`0i6b4 0Qv6`6I0f n T x ) s R P ) h  5 E R ) E ' e  ) k  5 D P 5 % D 4 # R ' 4 5 '  r r B  D P 5 % 5 E j R B r r q T T i R  5 c T R   ) T T @ R 22$Ii6ftP fS22U p32H hGR X`' 0ItUR X@ UGV h g   e d y u R q r r q T R   ) c T T @ R `IIfC2f62fFXvv2$$`' ItGU@ UUV h B R B B B R 7 @ R T T R 5 ' $iv2UY3$i22 $2I2) % c T T % R    ) 8 T T R    ) 8 T T R D D ) 4 E D  ) s T T y R ' 5 x  SR u b35 3GSFiR Sy S6`5 3GSFSR Sy 36I6G0IQSR Y 33``vF# w w R q B R @ u 8 s 2$2vvfUQQFt R r r q T h R E 4 1 e 5 @ c T T a R  5 D  % T @ R ) D ) V 1 ' D T T 7 R ' ' 2S23pigfG2H S`3dSR U@ b36`00YSR X330WG003R U@ S6I Q#

P D ) ' 5  E E D B @ 87 5 4  1 ) ' % #       2) 0G1 2I60H GFCA&9&60320(&$" !

258

ANNEXE C. ARTICLE THE NATURE OF IRS 13

Annexe D Article The Galactic Center source IRS 13 : a star cluster


Contribution pour la confrence Galactic Center workshop 2002, relative au travail sur IRS 13E.

259

260

ANNEXE D. ARTICLE THE GALACTIC CENTER SOURCE IRS 13

261

Astron. Nachr./AN 324, No. S1, 0 7 (2003) / DOI theDOIprex/theDOIsufx

The Galactic Center Source IRS 13E: a Star Cluster


Jean-Pierre Maillard1 , Thibaut Paumard1 , Susan Stolovy2, and Francois Rigaut3
1 2 3

Institut dAstrophysique de Paris (CNRS), 98b Blvd Arago, 75014, Paris, FRANCE SIRTF Science Center, CalTech, MS 220-6, Pasadena, CA 91125, USA Gemini North Headquarter, Hilo, HI 96720, USA

Received 15 November 2002, accepted 2 December 2002 Key words Galactic Center, star cluster, WR star, infrared, adaptive optics PACS 04A25 High spatial resolution, near-infrared observations of the Galactic Center source, close to Sgr A , known historically as IRS13, are presented. These observations include ground-based adaptive optics images in the H, K and L bands, HST-NICMOS observations in lters between 1.1 and 2.2 m, and spectro-imaging data in the He I 2.06 m line and the Br line. Analysis of all these data has made possible to resolve the main component, IRS 13E, in a cluster of seven individual stars within a projected diameter of 0.5 (0.02 pc), and to build their SED. The main sources, 13E1, 13E2, 13E3 (a binary), and 13E4, are hot stars of different nature. 13E2 and 13E4 are emission line stars. The spectral type of the various members goes from O5I to WR, including dusty WRs like IRS 21 (Tanner et al. 2002). All these sources have a common westward proper motion. Two weaker sources, 13E5 and 13E6, are also detected within the compact cluster, with 13E5 proposed as another dusty WR and 13E6 as a O5V star. An extended halo seen around the cluster, part of the mini-spiral of dust is particularly enhanced in the L band. It is interpreted as a contribution of the scattered light from the inner cluster and the thermal emission from the dust. IRS 13E is proposed to be the remaining core of a massive, young star cluster which was disrupted in the vicinity of Sgr A , and hence, the possible source of the young stars in the central parsec, from the helium stars to the S stars.

1 Introduction
In the early mapping works of the central parsecs, a spot named IRS 13, bright at all near-infrared wavelengths, was reported, approximately located 3.6 south-west of Sgr A . It was later resolved into two sources in the K band separated by 1.2 , IRS 13E and IRS 13W (Simon et al. 1990). From spectroscopic studies in the same band, IRS 13W was identied as a cool star (Krabbe et al. 1995) and IRS 13E as an emission line source with strong He I 2.058, 2.112 m, Br line and other Brackett lines up to Br12 (Genzel et al. 1996), typical of the helium stars present in the central parsec. The rst adaptive optics (AO) map of IRS 13E in the K band obtained on the CFH Telescope was published by Paumard et al. (2001), showing that the source resolved into two equally bright components 13E1 and 13E2, plus a third weaker component called 13E3. Since the spectra of IRS 13E did not have the same spatial resolution the identication of the associated spectral type was becoming subject to caution. In the centimetric domain, Zhao and Goss (1998) found IRS 13 as the brightest radio continuum source after Sgr A at the Galactic Center. The detection of a discrete X-ray source from CHANDRA at the position of IRS 13 (Baganoff et al. 2001) was another element making IRS 13 a source of special interest. The high resolution images of the central parsec currently obtained at various wavelengths in the infrared are giving the possibility to study in detail this peculiar Galactic Center source. A complete description of the present work can be found in the companion paper of Maillard et al. (2003).
c 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0004-6337/03/S101-0000 $ 17.50+.50/0


262

ANNEXE D. ARTICLE THE GALACTIC CENTER SOURCE IRS 13

J.P. Maillard, T. Paumard, S. Stolovy, and F. Rigaut: The IRS 13 cluster

2 New data on IRS 13


Calibrated ground-based AO data from several telescopes and space-based NICMOS data in the near infrared, all containing IRS 13 in their eld, plus some spectroscopic data, and the proper motions of the sources of the IRS 13 eld, have all been combined. 2.1 High-angular resolution data The AO data come from two different systems, Gemini North (Graves et al. 1998) for the H and the Kp (2.12 m, FWHM 0.41 m) bands, and ESO 3.6-m telescope (Cl enet et al. 2001) for the L band. Medium (M) and wide-band (W) lters, respectively centered at 1.1, 1.45, 1.60, 2.22 m (coded F110M, F145M, F160W and F222M) and two close narrow-band (N) lters (F187N centered on the 1.87 m Pa line and F190N) were used in observing the stars at the inner parsec of the Galactic Center with the NICMOS cameras on board HST. A small portion of 2.5 2.5 , roughly centered on IRS 13E, of the image from each lter was analyzed. All these high-resolution, multi-band images have provided the spectrophotometric information on the IRS 13 sources and its environment, from 1 to 4 m. The Gemini AO data were calibrated by linear interpolation based on one bright, hot star of the eld under study, from the calibrated NICMOS data, between the F160W and F190N photometry for the H band, F190N and F222M for the Kp band. The star detection and photometry was made with the StarFinder procedure (Diolaiti et al. 2000) for all the images. Specially for the AO images a deconvolution code called MCS (Magain et al. 1998) was applied. For the H and the Kp images the width of the synthetic PSF was equal to 0.040 and to 0.192 for the L-band image, i.e. a gain in resolution respectively of a factor 4.5 in H, 4.3 in Kp and 1.5 in L. 2.2 Spectroscopic data The only high-spatial resolution images giving spectroscopic information are contained in the NICMOS narrow-band images from the F187N lter. By subtraction of F190N, a narrow-band lter in the nearby continuum, from F187N, a map of the 1.87 m Pa emission was obtained (Stolovy et al. 1999). This map shows the distribution of the ionized gas and stellar spots from Pa emission in the atmosphere of the hot stars. The Br and 2.06 m He I line prole at IRS 13E from BEAR spectro-imagery, an imaging FTS (Maillard 2000), were used as complementary information to help precise the spectral type of the underlying stars. The IRS 13 complex is located in a region of intense interstellar emission. The data cube was particularly useful to correct the two emission line proles from the interstellar emission, leaving fully resolved stellar proles respectively at 21.3 and 52 km s resolution. 2.3 Proper motions The proper motions of the IRS 13 sources and the sources contained in the surrounding 2.5 2.5 were obtained from Ott et al. (2003) who conducted an analysis of ten years of SHARP data (Eckart et al. 1995), providing more than 1000 proper motions in the central parsec.

3 Results
From the deconvolution analysis of the 2.5 2.5 eld including IRS 13, 20 individual sources were identied. IRS 13E is decomposed in seven sources, respectively names 13E1, 13E2, 13E3A and B, 13E4, 13E5 and 13E6. The name 13E3A and B is proposed for the two components of the source 13E3 which appears double only after deconvolution, in the H and Kp bands. The positions of all the sources and their proper motions, estimated for most of them, are given in Fig. 1. Their observed photometry in the H, K and L bands is presented in Table 1. In the F187N - F190N image only 13E2 and 13E4 are remaining, indicative that these two sources are emission stars.


263

Astron. Nachr./AN 324, No. S1 (2003)

Fig. 1 The IRS 13 eld with the star identication. The vector associated with most of the stars represents the amplitude in velocity and the direction of proper motion measured from SHARP data by Ott et al. (2003). For E3A and E3B only the barycenter proper motion is determined. The origin of the eld corresponds to -2.19 W and -3.00 S with respect to Sgr A .

From the ux measurements made in 8 bands, between 1.1 and 3.5 m, it was possible to obtain a dereddened spectrophometry of the IRS 13 sources. That supposes to adopt a value and a law of dereddening. From a mean value of 31.1 mag is known to strongly vary with extinction within the central parsecs (Scoville et al. 2003). The local value was derived from two constraints: IRS 13W known as a cool oxygen star, and IRS 13E2 and 13E4 as hot stars from Pa imaging. A value of was adopted and was assumed to be valid over the small eld around IRS 13. could be pretty well constrained assuming that IRS 13E2 is a WR star with T 25,000 K. The IRS 13E2 spectral type is based on the Pa imaging, but also on Fabry-P spectro-imaging behind the CFHT-AO system where erot enet et al. 2003). IRS 13E2 is detected as the only source of the broad 2.06 m He I emission line (Cl With the spectral range under study, from about 1 to 4 m, beyond a temperature of 25,000 K we are in the Rayleigh-Jeans regime, and the shape of the SED becomes constant in a versus diagram. can be adjusted to bring the data points parallel to the SED. However, the t of the dereddened data has to be made as the sum of two black-body curves since most of the sources have an infrared excess, signature of thermal dust emission. The adjustments is made with four parameters for each IRS 13 star, by . The T temperature being the high-temperature component, is mainly determined by the data points between 1 and 2.5 m, and T by the 2 to 4 m points. If for the tting T becomes 25,000 K the temperature is set x. The four nal parameters are presented in Table 2. The dereddened points and the ts are shown on Fig. 2. Several stars are very hot stars, i.e. with

 

42 0( % # ! 531)' &$"

  

2V ( E E C VA9! 7 S 2B ( E E C BA9! XPIHWW@@@8UTRQPIHGFD@@87

264

ANNEXE D. ARTICLE THE GALACTIC CENTER SOURCE IRS 13

J.P. Maillard, T. Paumard, S. Stolovy, and F. Rigaut: The IRS 13 cluster

aT 25,000 K. The NICMOS F110M and the L data are essential to give the maximum of constraints to the SED of each source.

4 Nature of the IRS 13E sources


From the main results reported in the previous section an identication of the spectral type of the seven components of IRS 13E can be derived.
Table 1 H, K and L photometry of the IRS 13E cluster and the nearby eld stars
 

ID W E1 E2 E4 5 6 7 8 E3A 10 E3B 12 13 14 15 16 E5 18 E6 20

14.51 12.71 13.02 14.34 14.28 14.68 14.38 14.82 16.44 15.51 16.93 15.89 15.91 15.91 16.61 16.62 18.15 16.77 16.20 17.26

11.22 10.90 10.95 11.64 11.82 12.05 12.10 12.14 12.38 12.71 12.98 13.19 13.26 13.33 14.12 14.13 14.20 14.43 14.48 14.70

8.92 8.59 7.73

7.50 7.92

8.48

3.29 1.81 2.07 2.70 2.47 2.63 2.28 2.68 4.06 2.80 3.95 2.69 2.65 2.58 2.49 2.49 3.95 2.34 1.73 2.56

2.29 2.31 3.21

4.89 5.06

5.71

Table 2 Fitting parameters of the SED of the IRS 13E sources and IRS 13W The spectral type of each source as discussed in Sect. 5 is summarized in the last column.
   

a : dusty Wolf-Rayet star

Star W E1 E2 E4 E3A E3B E5 E6

Coef 23.00 0.500 0.450 0.070 0.460 0.375 0.070 0.008

T K 2600 25000 25000 25000 3800 3800 6000 25000

Coef 1700 12000 40000 45 33000 29000 9800

T K 650 550 550 1550 610 580 630

Sp. Type M5III O5I WC9 O5IIIe d. WR d. WR d. WR O5V

265

Astron. Nachr./AN 324, No. S1 (2003)

W E1 E2 E4 E3A E3B E5 E6

Fig. 2 Dereddened ux in W cm m for = 35. The top of the arrows represents the upper limit of the detectable ux in the 1.1 m and the L-band lters. The various lines represent the best tting between 1 and 4 m of the data points from a two-component model with the parameters of Table 2.

4.1 13E1, 13E2 and 13E4 The source 13E1 is a bright, blue star, but with no detected emission at Pa . From its luminosity and its IRS 13E1 is proposed to be close to a O5I main sequence star. 13E2 and 13E4 are two emission line stars, 13E2 being brighter in Pa by a factor 2. From the BEAR data, the 2.06 m He I line is a broad line ( 900 km s FWHM) while the Br line is narrow ( 215 km s FWHM). 13E2 is reported as the only He I emitter (Cl enet et al. 2003). As a broad-line, helium-rich star, 13E2 is proposed as a Wolf-Rayet type star, from the criterion on the linewidth developed in Paumard et al. (2001). By analogy with similar stars in the central parsec the source should be more precisely a WC9 star. 13E4 is a blue star which shows a narrow emission line in Pa but no helium emission line. Therefore, this star is much less evolved than 13E2. It can be proposed as a O5IIIe, since it is weaker than E1 and has hydrogen lines in emission.

 

 

  

266

ANNEXE D. ARTICLE THE GALACTIC CENTER SOURCE IRS 13

J.P. Maillard, T. Paumard, S. Stolovy, and F. Rigaut: The IRS 13 cluster

4.2 13E3A, 13E3B and 13E5 On Fig. 2 these three stars (dash-dotted lines) have a similar SED. They are adjusted by a strong, cool component at 600 K (Coef is high) and a weak, hotter component of a few thousands K. This adjustment can be compared to the tting of the SED of IRS 21 (Tanner et al. 2002) tted by a two-component model, the near-infrared scattered light from the central source peaking at 3.8 m (760 K), and the mid-infrared re-emitted light from a dust shell at 250 K. Tanner et al. (2003) conclude that this source is a dusty WR star, experiencing rapid mass loss, and the other red, featureless spectrum sources along the Northern Arm IRS 1W, 2, 3, 5, and 10W, as well. By analogy, we conclude that the three red sources within the IRS 13E complex, also located in the dusty part of the mini-spiral are dusty WR stars. 4.3 13E6 The source IRS 13E6 is another blue star, much weaker than 13E1, 13E2 and 13E4, with K 14.5 (Table 1). The image in the L band is not deep enough to detect it at this wavelength to conrm that this star is also embedded in the same concentration of dust than the other IRS 13E sources. From its color and magnitude IRS 13E6 can be considered as close to a O5V type star. Without further indication, only from the fact that all the other stars in IRS 13E are hot stars, we assume that IRS 13E6 belongs also to the same complex.

5 Model of IRS 13E as the remaining core of a massive star cluster


IRS 13E appears as only composed of hot, massive stars, with at least 7 stars within 0.5 . The common direction and comparable amplitude of the proper motions of the main components is a decisive argument to indicate that 13E1, 13E2, 13E3A/B and 13E4 are physically bounded. The source previously called IRS 13E is likely a compact star cluster. Furthermore, its composition means a young star cluster of a few 10 yr old, since several members are identied as having already reached the WR stage. The presence of such a compact cluster with a limited number of members raises the question of its origin. First, it can be noticed that each component has many other examples of stars of the same spectral type in the central parsecs. However, the large abundance of massive stars, which are very rare elsewhere in the Galaxy, remains one of the major mysteries of this region of the Milky Way. Since star formation would be difcult due to the strong tidal forces from the Sgr A black hole, Gerhard (2001) made the interesting hypothesis that the central parsec He I stars, the most prominent of the massive young stars, might be the remains of a dissolved young cluster, disrupted in the vicinity of the central black-hole. Kim et al. (2003) tested this idea for different cluster masses and different initial orbit radii. They came to the conclusion that some simulations can be regarded as possible candidates for the origin of the central parsec cluster. With its exceptional concentration of massive stars, very close to Sgr A , all bounded together, we propose that IRS 13E might be the remaining core of such a massive cluster which was disrupted by Sgr A . The analysis of the 12 other stars identied in the IRS 13 eld (Fig. 1) conducted as for the IRS 13E cluster sources, made possible to separate the sources in two categories, 9 red stars (T from 2800 to 5000 K) and 3 blue stars (T 25,000 K). The red stars are members of the most numerous population of the central parsecs, which is an old population of K, M and AGB stars, to which belongs also IRS 13W. The blue stars should be members of the most recent stellar population. On the other hand, the blue stars are comparable in magnitude and color to the stars of the S-cluster (Gezari et al. (2002) detected around Sgr A . The IRS 13E cluster itself contains also one of such lower mass blue stars (Table 2). Hence, the hot stars, including the S and the helium stars, could come from the same initial massive cluster, and complete its IMF. However, more simulations are needed to validate this hypothesis. Another aspect of IRS 13 is the detection of a discrete X-ray emission within 1 positional accuracy (Baganoff et al. 2001). IRS 13E as a star cluster might be the X-ray source, by the colliding winds of all the close, hot, mass-losing stars. An example of such a source can be provided by the detection of a

 





  

267

Astron. Nachr./AN 324, No. S1 (2003)

discrete X-ray source at the position of the core of the Arches cluster (Yusef-Zadeh et al. 2002). A better astrometry of the X-ray source at IRS 13 could help conrm this assumption, consistent with IRS 13E as the remaining core of a massive star cluster.

References
Baganoff, F.K., Maeda, Y., Morris, M., Bautz, M.W, Brandt, W.N., Cui, W., Doty, J.P., Feigelson, E.D., Garmire, G.P., Pravdo, S.H., Ricker, G.R., & Townsley, L.K. 2001, ApJ, astro-ph/01002151 Cl enet, Y., Rouan, D., Gendron, E., Montri, J., Rigaut, F., L P. & Lacombe, F. 2001, A&A, 376, 124 ena, Cl enet, Y., Lacombe, F., Gendron, E., & Rouan, D. 2003, these proceedings Diolaiti, E., Bendinelli, O., Bonaccini, D., Close, L., Currie, D. & Parmeggiani, G. 2000, A&AS, 147, 335 Eckart, A., Genzel, R., Hofmann, R., Sams, B.J., & Tacconi-Garman, L.E. 1995, ApJ, 445, L23 Genzel, R., Thatte, N., Krabbe, A., Kroker, H., & Tacconi-Garman, L.E. 1996, ApJ, 472, 153 Gezari, S., Ghez, A.M., Becklin, E.E., Larkin, J., McLean, I.S., Morris, M. 2002., ApJ, 576, 790 Gerhard, O. 2001, ApJ, 546, L39 Graves, J.E., Northcott, M.J., Roddier, F.J., Roddier, C.A., & Close, L.M. 1998, SPIE, 3353, 34 Krabbe, A., Genzel, R., Eckart, A., Najarro, F., Lutz, D., Cameron, M., Kroker, H., Tacconi-Garman, L.E., Thatte, N., Weitzel, L., Drapatz, S., Geballe, T., Sternberg, A., & Kudritzki, R.P. 1995, ApJ, 447, L95 Kim, S.S., Morris, M., & Figer, D.F. 2003, these proceedings Maillard, J.P., 2000, in Imaging the Universe in 3 Dimensions, ed. E. van Breughel & J. Bland-Hawthorn, ASP Conf. Ser., 195, 185 Maillard, J.P., Paumard, T., Stolovy, S., & Rigaut, F. 2003, A&A, submitted Magain, P., Courbin, F. & Sohy, S. 1998, ApJ, 494, 472 Ott, T., Genzel, R., Eckart, A., & Sch del, R. 2003, these proceedings o Paumard, T., Maillard, J.P., Morris, M., & Rigaut, F. 2001, A&A, 366,466 Simon, M., Chen, W.P., Forrest, W.J., Garnett, J.D., Longmore, A.J., Gauer, T., & Dixon, R.I. 1990, ApJ, 360, 95 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher, M., & Yusef-Zadeh, F. 2003, these proceedings Stolovy, S.R., McCarthy, D.W., Melia, F., Rieke, G., Rieke, M.J., & Yusef-Zadeh, F. 1999, in The Central Parsecs of the Galaxy, ed. H. Falcke, A. Cotera, W.J. Duschl, F. Melia, M.J. Rieke, ASP Conf. Ser., 186, 39 Tanner, A., Ghez, A.M, Morris, M., Becklin, E.E., Cotera, A., Ressler, M., Werner, M., & Wizinovitch, P. 2002, ApJ, 575, 860 Tanner, A., Ghez, A.M, Morris, M., & Becklin, E.E. 2003, these proceedings Yusef-Zadeh, F., Law, C., Wardle, M., Wang, Q.D., Fruscione, A., Lang, C.C., & Cotera, A. 2002, ApJ, 570, 665 Zhao, J.H., & Goss, W.M. 1998, ApJ, 499, L163

268

ANNEXE D. ARTICLE THE GALACTIC CENTER SOURCE IRS 13

Annexe E Article Kinematic and structural analysis of the Minispiral in the Galactic Center from BEAR spectro-imagery

269

270

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL

271
Astronomy & Astrophysics manuscript no. dynamicsnobf (DOI: will be inserted by hand later)
July 30, 2003

Kinematic and structural analysis of the Minispiral in the Galactic Center from BEAR spectro-imagery
T. Paumard1 , J.P. Maillard
1 2

1,

and M. Morris2

Institut dAstrophysique de Paris (CNRS), 98b Blvd Arago, 75014 Paris, France University of California, Los Angeles, Div. of Astronomy, Dept of Physics and Astronomy, Los Angeles, CA 90095-1562, USA

Received date / Accepted date


Abstract. Integral eld spectroscopy of the inner region of the Galactic Center, over a eld of roughly 40 40 was obtained

at 2.06 m (He I) and 2.16 m (Brackett-) using BEAR, an imaging Fourier Transform Spectrometer, at spectral resolutions respectively of 52.9 km s1 and 21.3 km s1 , and a spatial resolution of 0.5 . The analysis of the data was focused on the kinematics of the gas ows, traditionally called the Minispiral, concentrated in the neighborood of the central black hole, Sgr A . From the decomposition into several velocity components (up to four) of the line prole extracted at each point of the eld, velocity features were identied. Nine distinguishable structures are described: the standard Northern Arm, Eastern Arm, Bar, Western Arc, and ve additional, coherently-moving patches of gas. From this analysis, the Northern Arm appears not limited, as usually thought, to the bright, narrow North-South lane seen on intensity images, but it instead consists of a weak, continuous, triangular-shaped surface, drawn out into a narrow stream in the vicinity of Sgr A where it shows a strong velocity gradient, and a bright western rim. The Eastern Arm is split into three components. An Eastern Cavity proposed on radio maps is not conrmed, but a new ISM feature is detected just east of its position, the Eastern Bridge, that seems to be a new ow. We also report absorption of ISM structures by others, providing information on their relative position along the line of sight. A system of Keplerian orbits can be tted to most of the Northern Arm, and the bright rim of this feature can be interpreted in terms of line-of-sight orbit crowding as being formed by the warping of the owing surface at the western edge facing Sgr A . These results lead to a new picture of the gas structures in Sgr A West, in which large-scale gas ows and isolated gas patches coexist in the gravitational eld of the central Black Hole. The question of the origin of the ionized gas is addressed and a discussion of the lifetime of these features is presented.
Key words. infrared spectro-imaging FTS Galaxy: Center Sgr A West ionized gas

complex in the radio recombination H92 line at 3.6 mm (8.3 GHz), also at a resolution of 1 . Much higher spatial resoluWithin the inner 2 pc of the Galactic Center (GC) lies the Sgr A tion was reached with the VLA at 13 mm, with a beam size of West region, dominated by ionized gas which, because of high 0.15 0.10 , in the course of a project to measure proper moobscuration along the line of sight, has been detected only at in- tions of the bright, compact blobs of ionized gas (Zhao & Goss frared and radio wavelengths. Infrared ne-structure line emis- 1998). Also, with the NICMOS cameras on board HST, the sion of [Ne ] at 12.8 m has been used to map the gas dis- Pa line was observed at a spatial resolution of 0.18 (Scoville tribution a number of times, with successively higher spatial et al. 2003, these data will be used in this paper for comparison sampling and spatial and spectral resolutions, up to 0.6 1.0 purposes). sampling, 30 km s1 and 2 resolution in the most recent Br at 2.166 m has also been used to trace the ionized gas. paper (Lacy et al. 1991). In parallel, observations with the Very The rst detection consisted of a grid of spectra around IRS 16 Large Array (VLA) telescope provided a 6-cm map of the ion(Geballe et al. 1991) which could not give an overview of the ized gas in the radio continuum at 1 resolution (Lo & Claussen emission morphology. The availability of near-infrared arrays 1983). Later, Roberts & Goss (1993) observed the Sgr A West has resulted in many images of the Galactic Center. However, the ionized gas can only be detected by spectro-imaging or by Send oprint requests to: J.P. Maillard, e-mail: maillard@iap.fr. Data available online at CDS narrow-band imaging on a strong emission line. Broad-band images, for example in the K band, are dominated by the stel(http://cdsweb.u-strasbg.fr/), see Sect. 5.5. Visiting Astronomer, Canada-France-Hawaii Telescope, operated lar content. A rst attempt of spectro-imagery in Br was made by the National Research Council of Canada, le Centre National de la by Wright et al. (1989) with a Fabry-Perot system scanned over 1 000 km s1 , at a modest spectral resolution of 90 km s1 on Recherche Scientique de France and the University of Hawaii.

1. Introduction

272
2

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

a 38 36 eld. The data cube obtained in the same line with BEAR, an Imaging Fourier Transform Spectrometer on the Canada-France-Hawaii telescope represents a signicant eort to cover most of the central ionized region with a much better spectral resolution (FWHM 21.3 km s1 ), at seeing-limited resolution. A preliminary analysis was presented by Morris & Maillard (2000). Data from the same instrument were obtained on the 2.06 m He line, leading to the rst identication of interstellar Galactic center gas in this line (Paumard et al. 2001, hereafter Paper I). Data were also obtained with NIRSPEC on Keck II, by scanning the eld with the 24 slit used in a northsouth orientation to obtain a spectral cube covering 1.98 m to 2.28 m at resolution of 21.5 km s1 (Figer et al. 2000). All these data show that the ionized gas in the inner few parsecs of the Galactic Center is organized, in projection, into a spiral-like morphology having several apparent arms. This has led to the widespread appellation, Minispiral for this entire pattern. The brightest features are named Northern Arm, Eastern Arm, Bar, and Western Arc, as if imitating the morphology of a very small spiral galaxy. These terms seem to imply that the ionized lamentary structures constituting Sgr A West either form spiral patterns, or are portions of spiral arms. This view was motivated by the gas dynamical study carried out by Lacy et al. (1991), who interpreted the [Ne ] data in terms of a one-armed linear spiral in a Keplerian disk. The kinematics derived from all these data were also very useful in constraining the enclosed mass. The various features of Sgr A West give a spiral appearance primarily because of the way they are superposed on each other. However, a new analysis of Lacys data was conducted by Vollmer & Duschl (2000) to re-examine the kinematic structure of the ionized gas. Using a three-dimensional representation they conrm the standard features, but with a more complex structure, including two features for the Eastern Arm: a vertical nger of high density and a large ribbon extending to the east of Sgr A , and two distinctly dierent components in the Bar. Data in dierent lines, at better spectral and spatial resolution, warrant an independent kinematic analysis. The ionized gas is one component among the constituents that coexist within the central deep well of gravitational potential created by the black hole candidate (Sgr A ) of a mass equivalent to several million solar masses, with the compact cluster of young stars, the surrounding population of evolved stars and the ring of molecular gas. All these constituents orbit around the central dark mass. However, a consensus concerning its mass does not seem to have been emerged yet: assuming a distance of 8 kpc (Genzel et al. 2000), Ott et al. (2003) give a mass of 2.9 0.2 106 M , whereas Ghez et al. (2003) give 4 106 M . Developing a detailed picture of these components will improve our understanding of the interaction of all these components in the Galactic Center. The current analysis completes the work made from similar data on the central cluster of young, helium stars presented in Paper I and updated in Paumard et al. (2003). In the present paper, the gas content in the inner region of the GC is presented and analyzed from high spectral resolution data cubes on the Br and the He 2.06-m line, obtained with BEAR. The He data are from a new data cube (larger

eld, improved spectral resolution) compared to the data used in Paper I. A multi-component line tting procedure applied to the emission-line proles in each point of the eld is described in Sect. 3. It was used rst on the Br cube and then on the He cube. From this decomposition in Br, the identication of dened gas structures comprising the whole Sgr A West ionized region is presented in Sect. 4. A comparison between the decompositions in Br and He is presented in Sect.4.3. Attempts to adjust Keplerian orbits to the owing gas are presented in Sect. 5, which contains in Sect. 5.5 a discussion of the implication of these identications for the formation and the lifetime of the inner ionized gas.

2. Observations and preparatory data reduction


The 3-D data analyzed in the paper were obtained during two runs with the BEAR Imaging FTS (Maillard 1995, 2000) at the f/35 infrared focus of the 3.6-m CFH Telescope. In this mode, a 256256 HgCdTe facility camera is associated with the FTS, in which several narrow-band lters are selectable. Two of them were used, one which contains the Br line (4616.55 cm1 , bandpass 4585 4658 cm1 ) and the other one centered on the He line at 4859.08 cm1 (bandpass 4806 4906 cm1 ). The eld of view of the instrument is circular, with a diameter of 24. The Br data were acquired in July 25, 26, 1997 (UT) by observing two overlapping elds in order to cover most of a eld of 40 28 , oriented in the East-West direction, centered on the position of Sgr A (Fig. 1). The raw data consist of cubes of 512 planes with an integration time of 7 s per image. From the maximum path dierence which was reached, the corresponding limit of resolution (FWHM) in velocity is equal to 21.3 km s1 . On the following night a single eld centered on Sgr A was recorded with the 2.058 m He lter. The analysis of the later high resolution data was reported in Paper I, which brought new results on the central cluster of massive, hot stars, and led to the detection of the Minispiral in helium. However, the eld was not large enough for a signicant areal coverage of the Minispiral. New observations through the same lter were therefore obtained on June 9, 10, 11, 2000 in order to get three overlapping circular elds covering, when merged, most of a total eld of 36 36 , also centered on Sgr A . The estimated width of the interstellar 2.058-m line in Paper I called for an improved spectral resolution. A value of 50 km s1 (exactly FWHM 52.9 km s1 ) was chosen instead of 74 km s1 in the previous data, not as high as for Br, since the line is weaker. The raw data consist of cubes of 401 planes with an integration of 20 s per image, double the time for the previous data, to improve the detection depth. The processing of the BEAR data was presented in Paper I; the main steps are standard cube reduction, atmospheric OH correction and correction of lter transmission and telluric absorption particularly important for the 2.06-m data. The OH correction was more dicult for the Br data and for the new He data because of the higher spectral resolution and the better signal-to-noise ratio, making the OH lines stronger and the sidelobes of the prole more extended. The next step is the generation of the line cubes, spectral cubes in which the con-

273
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 3

and the Minicavity, are indicated. Also, a few emission line stars show up as bright points in the image.

is to separate these various ows and to describe them independently from each other. For this purpose, the development of a Minicavity multi-component line tting procedure able to work on 3D data appeared to be absolutely required. From a coarse examination of the datacube with cubeview, tting with a maximum of four distinct velocity components along each line of sight seemed SgrA* + adequate. Northern Arm Bar A comparison of the velocity components from one line of sight to the next should usually reveal coherent velocity structures by continuity. In the end, it might be possible to conclude whether these structures are isolated, or form continuous ows. Thus, the process is split into two main parts: rst the line prole decomposition at all the points of the eld, and second, Eastern Arm the structure identication. However, these two steps must be conducted iteratively in order to take full advantage of the 2D Fig. 1. Three color image of the two mosaicked elds of Sgr A West information in the cube. This work is based on original softobserved with BEAR in Br, between 350 (purple) and +350 km s1 ware developed by Miville-Desch nes (personal communicae (red). The standard bright features, Northern and Eastern Arms, Bar, tion), which we have largely extended.

3.1. Line prole decomposition


tinuum level in each point of the eld is tted and subtracted, in order to keep only the emission lines. The separation of stars and gas was not applied to the Br cube since the interstellar medium (ISM) dominates the Br emission. It had to be applied to the He data, as explained in Paper I. The merging of the line cubes from dierent data acquisitions required a new procedure to generate the nal line cube of the full eld. As an illustration, Fig. 1 was obtained from the Br merged cube with cubeview for Yorick, a port to the Yorick interpreted language of the facility program specially developed under IDL to examine the BEAR data cubes (Maillard 2000). The Br line cube is dominated by the emission from the interstellar medium (ISM), but a thorough inspection with cubeview shows that some stars exhibit the Br line in emission (Fig. 1). On the contrary, in the He line cube the stellar emission from the hot stars predominates (Paper I), but the ISM emission is clearly detected too, with a better contrast, thanks to the higher spectral resolution. The central parsec was observed with NICMOS cameras onboard HST, during a few runs between Aug. 1997 and Aug. 1998, in 6 near-IR lters, including 2 narrow-band lters, F187N centered on 1.87 m Pa, and F190N, the nearby continuum. By subtracting the F190N lter from the other one, Pa emission was obtained on a eld of 19 19 centered on Sgr A (Stolovy 1999, Scoville et al. 2003) at a spatial resolution of 0.18 , and a wider eld of 120 at a lower resolution of 0.4 . We use an image covering the central 40 40 eld from these data for the purpose of comparison, in Fig. 12.

3.1.1. Line prole


A single velocity component of the emission lines from the ISM has been assumed to be Gaussian, given as a function of v by : vv0 2 I(v) = I0 exp 2 where I0 is the amplitude of the Gaussian expressed in erg s1 cm pixel1 (1 pixel = 0.125 arcsec2 ), v0 is the radial velocity of the component, and the width of the line, due to thermal agitation, turbulence and any velocity gradient along the line of sight. The total ux per pixel of the line is then: = 2 I0 0 c

where c is the velocity of light and 0 the rest wavenumber of the studied line. The full width at half maximum (FWHM) of the line is given by: FWHM = 2 2 ln 2 The detected spectrum is convolved by the instrumental line shape (ILS) of the FTS, which is by denition a sinc function dened by: sin m v0 c (v) = m v0 c

where m is the maximum path dierence between the two arms of the interferometer, that determines the limit of resolution d At each point of the eld the Br emission prole generally of the data with d = 0.6/m (FWHM). appears complex. The basic assumption which is made is that The measured line prole is thus the convolution product each observed prole results from the combination of several S = I , function of three free parameters I0 , and v0 . Each velocity components, that is, that along any given line of sight single spectrum of the eld has been tted to a set of four such several ows are superposed. The rst goal of the present paper lines, thus implying twelve free parameters.

3. Structure identications

274
4

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral Table 1. Common names of the narrow (N) and broad (B) line stars drawn on Fig 3 (see Paumard et al. 2003, and references therein). ID N1 N2 N3 N4 N5 N7 Name IRS 16NE IRS 16C IRS 16SW IRS 16NW IRS 33SE IRS 34W ID B1 B2 B3 B4 B5 B6 Name ID 180 IRS 7E2 IRS 9W IRS 15SW IRS 13E2 IRS 7W ID B7 B8 B9 B10 B11 B12 B13 Name AF AFNW HeIN3 BSD WC9 IRS 29N IRS 15NE IRS 16SE2

3.1.2. Procedure
The tting routines use MPFIT , a general purpose tting engine written in IDL. It has been prefered to the standard IDL tting procedure because of its greater robustness and because of its versatile interface that allows one to set complex constraints on the variables. The whole procedure is divided into several steps, detailed below. Preparation: As for any tting routine, a reasonable initial guess must be provided for each point of the eld. For such a problem, where we intend to t complex line proles at low signal-to-noise, the method cannot be fully automatic. The method consists of determining an initial guess only for a few points, and letting the software determine initial guesses for the other points from these, as we will see in the next paragraph. Good initial guess are chosen for a few starting points in the eld, selected by the user for their high signal-to-noise associated line proles and unambiguous decomposition. The user decides the optimum number of starting points and their locations. However, they should be chosen so that every structure in the eld is represented, and the most complex regions are better tted if they are close to a starting point. In our case, ve starting points were used. Step 1: From the starting points, a rst procedure attempts to t a four-component line shape function to each spectrum. For each new spectrum, the initial guess is determined from the results found for the neighboring points. The spectra are studied sequentially in parallel spiral-mode scannings around each starting point. Except for the initial guess, the tting of a spectrum is independent of all the others. Step 2: The velocity structures are then built. For the brightest point of the eld, the neighbors are examined, and searched for a component such that the velocity gradient between the point of interest and this neighbor is less than a certain amount, which is set by the user at runtime. The procedure is iterative, and once a few neighbors have been selected into a structure, their neighbors are in turn examined for possible selection. The procedure stops when every component of every point of the eld has been assigned to exactly one spatial structure. This procedure allows only one component of a given point to be selected into a given structure. A structure that overlaps itself spatially, thus causing two velocity components on the same line of sight, cannot be directly detected as such: the program splits it into two structures. Step 3: This procedure requires that the detected structures be manually inspected. The user has then the possibility to add some more common sense heuristics into the structure identication, a little dicult to implement but easy to apply manually. Several problems can occur: during step 1, the tting procedure might t only one component where two blended components are indeed more appropriate, during step 2, if two overlapping structures intersect each other in the velocity space, the procedure can
1

falsely cross-connect them, i.e. reconstruct two structures, each one being made of parts of both physical structures (Fig. 2). Step 4: Then, these manually corrected results are used to perform a second t at each point of the eld; at this point, 2D information is entirely included in the initial guess provided to the tting procedure. Iteration: Steps 2, 3 and 4 must be iterated a number of times, until a stable set of plausible structures is reached. Plausible structures means only that the structures are more extended than the spatial resolution, and brighter than the detection limit of the data.

4. Results

4.1. General description of the results


The analysis described above leads to a vision of the Minispiral more complex than usually thought, one which is consistent with, but more detailed than the description proposed by Vollmer & Duschl (2000). After a careful examination we identify 9 components of various sizes, labeled (a) to (i). The radial velocity maps of the structures are presented in Appendix A, and their ux maps in Appendix B. Two types of velocity map appear, some with a signicant overall velocity gradient, others without any appreciable, large-scale velocity gradient. The deviation from mean motion, dened as the local dierence between the velocity measured at one point and the mean value for the neighboring points, and divided by the uncertainty, ranges from roughly one tenth to ten for all the features, which means that every velocity structure shows signicant (over 3 ) local features. The areal size of the structures (Table 2), expressed in terms of solid angle covered on the sky, ranges from 17 arcsec2 to 300 arcsec2 for the part of the Northern Arm that is visible in the BEAR eld of view. The surface area of each structure must be considered as a lower limit because BEAR may not detect the weakest parts nor parts where blending with a brighter structure in the spectral domain prevents detection, and because the eld of view does not cover the entire Minispiral.

4.2. Morphology of the ionized gas in Sgr A West


A brief description follows for each identied velocity structure whose velocity maps are given in Figs A.1 to A.9. Table 2 gives the surface coverage on the sky (within the BEAR eld

http://cow.physics.wisc.edu/craigm/idl/tting.html

275
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 5

velocity

velocity

d cos

d cos

Fig. 2. Cross-connection problem. Left: two physical structures (solid and dashed polygons) intersect each other in the velocity space. Right: the structures reconstructed by the software can be erroneous.

Northern Arm

Northern Arm Chunk

Western Arc

10 5 0

Bar Overlay

Eastern Bridge
5 10

Western Bridge

20

15

10

10

15

Eastern Arm

Tip

Bar

Fig. 3. Left: outline of every structure. The region lled in black corresponds to the points where two lines associated with the Northern Arm are detected (see text). Filled diamonds (resp. empty squares) represent narrow (resp. broad) line helium stars. Right: IDs of the helium stars (adapted from Paumard et al. 2003). See Table 1 for common names identication.

of view), and the maximum and minimum velocity within the given structure.

IRS 16SW (N3) and IRS 33SE (N5, Fig. 3). It extends further away on 5 to the north-west, and contains the pointlike feature just above the aperture of the Minicavity. On the few pixels where both features are detected, the secondary a) Northern Arm: Contrary to its standard description, the layer is 5080 km s1 more blueshifted than the main one. Northern Arm is not seen here as a bright N-S lane, but On the velocity maps, both layers are drawn, indicating the as an extended, triangular surface. One edge of this triangle velocity of the secondary one for the few points were both is the bright rim generally noticed, but it extends all the way are detected. The ux map gives the sum of the two layers. over to the Eastern Arm. The third edge of the triangle is the The direction of the Northern Arm motion (from north to edge of the eld, so viewing this feature on a larger eld south) has been established by Yusef-Zadeh et al. (1998). may yield a somewhat dierent description. As it reaches The kinematics of the Northern Arm will be thoroughly the Minicavity, the Northern Arm is split into two layers in studied in Sect. 5. the spectral direction. Both layers are clearly detected only for a few adjacent pixels (14), that correspond to the small b) Bar: The Bar is the most complex region, where at least three components are superimposed. The most important nger-looking feature north-east of the Minicavity. The feature is very extended from the Ribbon of the Eastern main layer contains all the Minicavity, while the second Arm (c) to the Western Arc (e) is very straight, and shows layer seems to be deected northward of the Minicavity, a smooth overall velocity gradient. Vollmer & Duschl and forms the small nger between the two helium stars

276
6

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

Table 2. Feature identications, with surface areas (pixels and square arcseconds), and minimum and maximum radial velocities (km s1 ). ID a b c d e f g h i Feature name Northern Arm Bar Ribbon Eastern Bridge Western Arc Western Bridge Tip Northern Arm Chunk Bar Overlay S (pix) 2414 1389 833 670 471 327 207 185 136 S (arcsec2 ) 300.8 173.1 103.8 83.5 58.7 40.7 25.8 23.1 16.9 Vmin -286.9 -211 132.9 34.9 -37.1 -121.1 222.9 14.9 -267.0 Vmax 188.9 196.9 242.9 182.9 74.9 100.9 339.0 74.9 -7.1

(2000) mention two complementary components of the Bar, which they call Bar 1 and Bar 2, though their description is not sucient to determine precisely the positions of these two suggested components. We see two additional features, which we propose to call the Western Bridge (f) and Bar Overlay (i). Parts of the Bar are also superimposed on almost every other structure, including the Ribbon of the Eastern Arm (c), the Tip (g), the Eastern Bridge (d) and the Northern Arm (a). c) Ribbon: As already described by Vollmer & Duschl (2000), the Eastern Arm region is split into two parts: a Ribbon and a Tip (g). The velocity gradient of the Ribbon is directed along the minor axis of the structure, not along its major axis as expected for a ow. d) Eastern Bridge: A structure of medium size extends from the Ribbon (c) to the bright rim of the Northern Arm. It does not show any large-scale velocity gradient, and its shape does not show any principal axis that would indicate a ow. It is superimposed on the faint regions of the Northern Arm, and partly superimposed on the Ribbon, the Bar and the Tip. Its southern side is parallel to, as well as superimposed upon, the Ribbon; the two structures are probably related, although their relative velocities dier by more than 50 km s1 . The name we propose is based on the fact that it lies between the two Arms of the Minispiral, both in the spatial and spectral dimensions, being hard to distinguish in the spectral dimension from the Ribbon on its southern side and from the Northern Arm on its northern side. It is also inspired by the fact that the most luminous part of it in our eld is a small vertical bar, seemingly connecting the bright parts of the Northern and Eastern Arms. However, the Pa map (Fig. 12) shows that this bar may extend outside our eld-of-view into an elongated feature parallel to the Ribbon. The lack of an overall gradient in the velocity map suggests that this feature is not much affected by shear. e) Western Arc: The Western Arc lies just at the edge of the eld, so we have access only to its innermost part. It is seen as a rather simple feature, with large scale velocity gradient. It is superimposed on the Western Bridge on a few pixels. The velocity eld that we measure is basically in good agreement with that of Lacy et al. (1991). f) Western Bridge: The Western Bridge is a tenuous, elongated feature oriented east-west and extending from the Bar to

the Western Arc. This structure, as well as the Bar and the Bar Overlay (i) upon which it is superimposed, contains in projection the helium star IRS 34W (N7, Fig. 3. g) Tip: The Tip is, in projection, a very concentrated and relatively small object with the most redward velocity in the region ( 300 km s1 ). The Tip has already been noticed by Vollmer & Duschl (2000) only on a morphological basis, as a nger-looking feature of the Eastern Arm in their three dimensional data. Here, we see that the Ribbon and the Tip are two distinct features, superimposed on the line of sight, thus we do not adopt the representation-dependent denomination Finger. At the elbow between the Ribbon and the Tip, in the IRS 9W region, is a bubble-like feature, or a Microcavity (radius 1 ), with a rather bright rim (Fig. 4), which appears at a specic velocity (230 km s1 ).

d (arcsec)

7 8 9 10 8 7 6 5 4 3 2

d . cos() (arcsec)
Fig. 4. Microcavity feature in the region of IRS 9W, represented by the black cross. Axes are osets from Sgr A . Integrated velocity range: 220240 km s1 .

h) Northern Arm Chunk: A small tenuous structure is seen superimposed on the Northern Arm, a few arcseconds north of IRS 7. It lies at the edge of our eld, so it could extend further out; however the Pa image shows a small, horizontal bar at its location, crossing the bright rim of the Northern Arm, that does not seem to be much extended.

277
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 7

i) Bar Overlay: The Bar Overlay looks like a small cloud that is superimposed upon the western region of the Bar and that shows a velocity gradient similar to the one of the main Bar at the same location, with an oset of 40 km s1 . It may indicate that these two features are closely related. They could, for example, be the two faces of a single neutral cloud, ionized by two distinct UV sources. In our nomenclature, no feature is named Eastern Arm . This is because the region is far too complex, and calling one of the feature by this historical name would lead to confusion, thus we reserve this name to the entire region, which indeed comprises the Ribbon, the Tip, and the Eastern Bridge. As already discussed, the Tip and the Ribbon seem to be two parts of an entity, separated by the Microcavity. The Eastern Bridge seems related to this entity, it may belong to it, or be interacting with it.

Table 3. [He ]/[Br] for the dierent structures, relative to the mean value <[He ]/[Br]>. The Minicavity is separated from the Northern Arm, as it warrants special attention. ID a b c d e f g h i
a b

Feature name Northern Armb Minicavity Bar Ribbon Eastern Bridge Western Arc Western Bridge Tip Northern Arm Chunk Bar Overlay

[He ]/[Br]a 0.74 0.85 0.99 0.78 1.09 0.52 1.73 2.64 1.81

normalized to its mean value except Minicavity

10 The same work of decomposition into velocity structures has been performed on the He data. It was more dicult than for 5 the Br data since the spectral resolution and signal-to-noise ratio are lower. The fact that the He data are dominated by the 0 emission from the helium stars also contributes to the greater complexity of this task. Thus we skipped the rst step of the decomposition process, and provided directly a complete set of 5 initial guesses based on the Br results, since at rst sight the distribution of ionized gas is globally the same in He . This method prevents the He analysis from being fully indepen- 10 dent, although steps 2, 3 and 4 were performed eight times, until the procedure converged satisfactorily. 20 15 10 5 0 5 10 15 The Minispiral is detected, as are all of the individual strucFig. 5. Comparison between integrated ux in Br (grey scale) and tures, except the Northern Arm Chunk. However, as will be detailed below, several dierences in the appearance of these He (empty contours). Axes are osets from Sgr A in arcsec. structures are noticeable. In order to quantify these dierences, we have built [He ]/[Br] line ratio maps for each structure, Northern Arm are not detected in He , this value may be normalized to the areal mean for this ratio over the union of even smaller. The line ratio is higher on the western side of all the structures. The [He ]/[Br] line ratio varies considerthe bright rim, and this rim has in He the shape of a part ably across the eld, so that, for instance the Northern Arm of a circle surrounding the IRS 16 cluster. This circle conbright rim and the Minicavity do not show the same shape in tinues further to the northwest, forming a rather faint horn He and Br (Fig. 5). These dierences will be detailed later. at the location where, in Br, the rim bends abruptly (5 to Table 3 shows the mean normalized [He ]/[Br] line ratio for the north and 5 to the east of Sgr A , spot C on Fig. 12). the dierent structures. It appears that this ratio is lower than The open ring of ionized gas surrounding the Minicavity the mean value for the main, well known features the Northern is on average brighter in He then the rest of the Northern Arm, The Ribbon, the Bar and the Western Arc, and higher for Arm relatively to the intensity distribution in Br. Its inthe smaller features. However, the values are computed only for nermost border is even brighter. Its western edge, where the features detected in both Br and He , so they do not take IRS 13 and IRS 2 lie, is very bright, and looks like a vertiinto account the faintest, least excited regions of each feature. cal bar going from IRS 13 almost to the declination of the Below are summarized the most noticeable dierences between AF star, making the Minicavity look angular. the He and Br images of the Minispiral structures. b) The Bar is the main feature with the highest [He ]/[Br] a) Though the Northern Arm remains the most prominent fearatio, with a normalized value of 0.99. However, we do not ture of the Minispiral, the mean value of its normalized detect helium towards the full extent of its Br counterpart. [He ]/[Br] line ratio (Minicavity excluded) is one of the d) The Eastern Bridge (Fig. 6) is clearly identied, but it lowest, with a value of 0.74, being only higher than presents a shape much dierent from the one observed the value measured for the small part of the Western Arc in Br. It is brighter on its southern side, and the norththat we detect. Considering that the faintest parts of the ern parts are not detected by the procedure. The southern

4.3. Comparison with He data

278
8

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

the Eastern Bridge contains a substantial amount of dust, responsible of the absorption of about 50% of the Br ux of the Northern Arm. The ux map of the Bar is less smooth than that of the Northern Arm, making such eects more dicult to see. However, on this map, the characteristic shape of the Minicavity is clearly identied in absorption. Again, that tells us that the Bar is behind the Northern Arm on the line of sight.

5. Keplerian orbit tting


The velocity maps show a view of the features very dierent from the usual ux maps which, by themselves, can be misleading. For instance the morphology of the Northern Arm with its typical bright rim may lead one to think of this rim as the true path for most of the material. On the other hand, the velocity map shows no peculiar feature at the location of the rim. This is particularly intriguing for the location where it appears to bend abruptly, just a few arcseconds north of IRS 1 and east of IRS 7 (spot C on Fig. 12). Thus we are led to the idea that the kinematics of the Northern Arm should be studied independently of its intensity distribution.

Fig. 6. Composite image of the Eastern Bridge, from images integrated from 50 to 100 km s1 in the Br (red) and He (blue) cubes.

As a rst attempt at using the information contained in the new tools that are the velocity maps, we tried to analyse the Northern Arm as a Keplerian system. For a rst, simple approach, we created a dedicated IDL graphical package called GuiMapOverlay (Fig. 7). With this tool, the user can easily adjust one Keplerian orbit over a velocity map, the location and mass of the central object being those of Sgr A (position of Sgr A relative to IRS 7 from Menten et al. 1997, position of IRS 7 relative to the other stars of the eld from Ott et al. 1999 and a distance of 8 kpc). The central mass is still a matter of debate. We have used a value of 3 106 M (Genzel et al. 2000, Ott et al. 2003 give a value of 2.9 106 M ) for most of our models. We will discuss the impact of changing this mass later. A Keplerian orbit in 3D is dened by ve orbital parameters: the eccentricity, two angles dening the orientation of the orbital plane, the periapse (distance of closest approach to the center of motion), and a third angle dening the position of the 4.4. Absorption by the structures periapse. One of the most dicult questions concerning the ionized feaOnce the user is almost satised with the orbital parametures is to nd out their relative positions, i.e. when two struc- ters found by trial and error, an automatic tting procedure can tures are overlapping, which one is closer to the observer. On be called. It is possible to x parameters, and the orbit can be two occasions, the ux maps (Appendix B) can be used to nd forced to go through a selected constraint point by tying the this information. periapse to the other parameters. After a few experiments with The ux map of the Northern Arm shows a region of low this tool, we are led to some general conclusions: intensity, which north-western boundary is a well dened line, good agreement can be found between observed and calcuapproximately north-east/south-west. This limit between a related velocities, except in the region of the Minicavity; gion of low intensity and a region of high intensity is most this model alone is not sucient to decide whether the orobvious south of IRS 1 and west of the Minicavity. This line bits are bound or not, or whether the data are compatible follows very closely the outline of the Eastern Arm. This gives with elliptical, parabolic, or hyperbolic motion. us two pieces of information: the Northern Arm is behind the Eastern Bridge on the line of sight; The second of these points is not satisfactory, as one of the most interesting questions is to decide whether the gas is bound.

parts extend horizontally, following the edge of the Eastern Arm Ribbon upon which it is superimposed, with a velocity oset between the Eastern Bridge and the Ribbon of about -50 km s1 (measured in Br, but the agreement in good between the two lines), which again suggests that the two features are related. The bow-shaped bright rim of the structure, which is almost vertical and give its name to the Eastern Bridge, is oset by about 1 towards west in He relatively to Br. e) A small part of the Western Arc is detected within our eld; its [He ]/[Br] value is the smallest, but only a few points are detected both in He and Br. g) Due to the lower spectral resolution of the He data (52.9 km s1 , vs. 21.3 km s1 in Br), the Tip is not separated from the Ribbon by our procedure in this band. However, it is clearly seen. It is the brightest feature relative to its Br counterpart, with a normalized line ratio of 2.64. The line ratio is also noticeably brighter on its southwestern edge than on its northeastern edge, which is not detected in He by the decomposition procedure. The Microcavity is also observed in He .

5.1. Fitting one orbit on a velocity map

279
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 9

tricity. To avoid studying only local minima in the parameter space, it is also important to use several initial guesses.

5.3. Homothetic hypothesis


We designed a tool quite similar to GuiMapOverlay to easily study whether the data are consistent with a homothetic2 set of orbits, which is the simplest model. With the hypothesis of a homothetic set of orbits, the eccentricity still cannot be well constrained. Bound orbits seem to be preferred, but the agreement is as good with circular orbits and very eccentric orbits, close to parabolic. The residual map always has the same shape: the observed velocities are always smaller than the computed ones along the inner edge of the bundle of orbits, and higher along the outer edge. The global agreement is always poor, with < 2>1/2 70.

A few of these homothetic models have been chosen as initial guesses for other adjustments, with released constraints. It Nevertheless, the rst point convinced us to carry on Keplerian is rst interesting to check the coplanar hypothesis, in which modeling. So, we attempted to model the Northern Arm with only the two parameters that dene the orbital plane are kept uniform, and the uniform eccentricity hypothesis. The agreeseveral orbits instead of only one. ment is much better when making either the eccentricity or the orbital plane free. In the following, both parameters are free. 5.2. Fitting a bundle of orbits on a velocity map Even with the most general situation, the parameters are still not constrained enough to decide whether the orbits are To t several orbits at a time on a velocity map, the dierall bound or not, to extrapolate the model outside the eld of ent orbits must be forced to be actually dierent. That can be view, nor even to derive reliably the direction of proper motion. done by forcing each one to pass through a dierent constraint However the models share a few characteristics that we judge point, as explained in the one orbit case. Obviously, the conto be robust because of their repeatability: straint points themselves must be chosen on dierent physical orbits. We have chosen to take the constraint points aligned 1. the orbital planes are close to that of the CND; across the gas lane, evenly spaced in projection on a line that 2. the orbits are not quite coplanar; the two angles that dene passes through Sgr A . We have tried two dierent such conthe orbital plane vary over a 10 range; straint lines (Fig. A.1), and found consistent results. The result 3. the eccentricity varies from one orbit to another, being close described here uses 50 constraint points, evenly spaced on the to parabolic or above for the innermost orbits, and closer to solid line of Fig. A.1. Each constraint point is given an index, circular (below 0.5) for the outermost. increasing from the point nearest Sgr A outwards, that is used to refer to a given orbit. To ensure a smooth model we are interested only in the 5.5. 3D morphology and time-scale of the Northern Arm global motion the four functions that map each constraint point to one of the parameters have been chosen to be deWe present here our best model, i.e., the one with the lowscribed as spline functions, uniquely dened by their value at a est < 2 >1/2 among the realistic models that cover most of number of control points, chosen among the constraint points. the Northern Arm. The laws used for this model are shown The number of points used to dene the spline function can be in Fig. 8. The agreement between the radial velocity map freely chosen to set the spatial resolution of the model. After of this model and the observed velocity map is good: < several attempts, we have chosen to x this number to four in 2 >1/2 = 26. The method is unbiased, and the mean error our nal model. Thus, having four functions (one for each of in the radial velocity from this model is 10 km s1 , as estithe orbital parameters), each of them being dened by four valmated from Fig. 9. A 3D velocity map of the Northern Arm ues, the model depends on sixteen parameters. from this model is available online in FITS format at CDS We designed a tting procedure to adjust this model based (http://cdsweb.u-strasbg.fr/). It consists of one map 2 on the observed velocity map by minimizing the reduced < >. for each 3D component of the velocity in km s1 (Fig. 10), and Thus, it becomes possible to either x some parameters, or to 2 force them to have the same value for each orbit. This way, for Two orbits are said to be homothetic when they are identical example, it is possible to check whether the observed velocity except for their scale, i.e. when they share the same orbital parameters, map is consistent with coplanar orbits or with uniform eccen- except the periapse distance.

Fig. 7. GuiMapOverlay: given a set of orbital parameters, this tool draws the path of the corresponding Keplerian orbit on a velocity map, extracts the observed velocities along this path, and plots both the observed and computed velocity curves. The < 2> is also computed and shown.

5.4. General case

280
10

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

Fig. 8. These ve plots show the laws of each of the ve orbital parameters in our best model. All laws are cubic spline functions of the orbit index (see text), except the periapse, which is derived from the other parameters to bind each orbit to its constraint point. The lled region corresponds to the 1 error bars. The errors are always much smaller for the outermost orbits than for the innermost, because less points are available for determination of the shorter, inner trajectories.

10 5 0 5 10

d (arcsec)

20
Fig. 9. Histogram of the radial velocity dierence between our model and measurement; it is close to a Gaussian distribution centered on zero with = 10 km s1 . That means that the method is unbiased, and that the mean error is 10 km s1 .

15

10

10

15

d . cos() (arcsec)
2 2 6 10

dd (equivalent arcsec)
Fig. 11. Elevation map for the detected part of the Northern Arm from our model, given in equivalent arcseconds for homogeneity. dd stands for dierential of the distance, positive dd means further away from the observer than the center of mass, Sgr A , at d = 0, d = 0, dd = 0.

one map giving the distance to the observer for each point of the eld (oset from center of mass in equivalent arcseconds, Fig. 11). The variations of the orbital parameters induce a particular 3D shape for the Northern Arm (Fig. 12): for all the noncoplanar models, the Northern Arm looks like a warped surface, and this warping induces a crowding of orbits that closely follows the bright rim of the structure. That suggests that the Northern Arm is either a warped planar structure, or the ionized surface of a neutral cloud. The bright rim itself is not only due to the stronger UV eld and a real local enhancement of the density, but also to an enhancement of the column-density due to the warping. An interesting point is that, in some models, no orbit follows the bright rim, which emphasizes that it is

really important to consider the dynamics independently from the morphology of the Northern Arm. Another characteristic present in all the models is that the period of the orbits ranges from a few 104 years to a few 105 years, which implies that the Northern Arm would have a completely dierent shape in a few 104 years, and cannot be much older than that time-scale. Since the agreement in radial velocity is now rather good, it makes sense to look at the deviations from global motion by

281
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 11

10

5
500 km s1

d (arcsec)

10

20

15

10

10

15

d . cos() (arcsec)

220

120

20

80

Radial Velocity (VLSR, km . s1)


Fig. 10. 3D velocity map for the Northern Arm from our model. The arrows show the derived tangential velocities.

looking at the extended features on the residual velocity map (Fig. 12): A) the ow shows a rather signicant deviation in the region just southwest of the embedded star, IRS 1W; this perturbation could be due to the interaction with this stars wind; B) the region of this model closest to the Minicavity is perturbed; C) another deviation is seen at the precise location where the bright rim bends abruptly, just east of IRS 7E2; D) nally, an elongated feature is seen on the fainter rim coming from IRS 1W towards the northeast.

ible with elliptical (bound) motion anymore. Thus, the exact value of this parameter is very important to conrm whether or not some of the Northern Arm material is on unbound orbits.

5.6. The Minicavity

The Minicavity is seen as low intensity level in the Northern Arm ux map. However, material of this structure is detected on the line of sight of this region. Even at that location, mostly devoided of ionized gas, the velocity map of the host structure is essentially that of a ow. However, though it is not obvious on the velocity map of the Northern Arm, a perturThis best model uses a central mass of 3 106 M . It bation of the motion is detected there. Our Keplerian model has been used as the initial guess for a nal adjustment using does not pass through the Minicavity, but just south of it. 4 106 M (Ghez et al. 2003). With this value of the central However, this model does not include the outermost regions mass, every remark made above is still valid. The < 2> is not of the Northern Arm. These regions are those with the most changed signicantly, so that this work cannot help measuring circular motion, that is essentially east-west within our eld. this mass. The most noticeable dierence is that the error bar So the material from these outermost regions may well pass for the innermost orbit is two times smaller, and is not compat- closer to the Minicavity, or even through its projected location.

282
12

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

emission from IRS 13E, clearly detected both in He (Paumard et al. 2001), but also in Br (Maillard et al. 2003). However, the ux map of the Bar peaks sharply on the same line of sight. This seems to show that the compact star cluster IRS 13E excites locally material in the Bar, and must be either embedded in it, or very close to it. However, coincidence between the bright spot at this location and the end of the western edge of the Minicavity seems to be a projection eect, and nothing physical. The last fact worth noting concerning the Minicavity is the small nger-looking feature at the north of its eastern side. As already discussed Sect. 4.2, two velocity components are detected at this feature. Both connect continuously with the Northern Arm in velocity. From their velocity, it seems that this nger is deected by the Minicavity towards the north and towards the observer.

6. Discussion
The geometry of the Northern Arm has been studied from its velocity map, leading to the conclusion that it may not be a planar structure, but rather a three-dimensional structure. Fig. 12a is quite compatible with the Northern Arm indeed being the ionized surface of a neutral cloud (as suggested by Jackson et al. 1993; Telesco et al. 1996). This gure also suggests that the Northern Arm and the Western Arc may be two parts of the same physical structure. The velocity derived from the model agrees with the measured velocity of the Western Arc with 50 km s1 (which is reasonable since its an extrapolation) and has the right gradient. However, this coincidence is lost when a central mass of 4 106 M is used. However, the adjustment is not made over the entire Northern Arm, since our eld of view is limited. The same study on a complete map of the Northern Arm would probably reveal whether it is bound, and whether the Northern Arm and Western Arc are one same physical feature. The tangential velocity eld of the Northen Arm (Fig. 10) is interestingly similar the magnetic eld derived by Aitken et al. (1998). We agree with them that this seems to imply that even if the Northern Arm is part of a cloud stretched by the tidal forces from Sgr A , its dynamics is suciently Keplerian for the magnetic eld lines to get aligned with the velocity. We have shown that at least two structures are thick, dusty clouds, because their absorption factor is of the order of several 10% at 2 m: the Eastern Bridge and the edges of the Minicavity in the Northern Arm. From the wider eld in Pa, we can assume that the Eastern Bridge is an elongated cloud, of which we detect only the western front, that seems to be forward front if its motion is mostly east-west. The lack of shear inferred from its velocity map can be explained by the fact that if it is really the forward front of the cloud, then the velocity eld must be perpendicular to its surface. The presence of three isolated ionized gas structures (the Western Bridge, the Northern Arm Chunk and the Bar Overlay) in addition to the standard large ows and to the Eastern Bridge, that seems to be another ow, has been demonstrated. Some of these structures may be isolated gas patches, but it is also possible that some of them are regions of the neutral clouds

Fig. 12. On this Pa map (Scoville et al. 2003), one of the Keplerian models is overplotted (top panel). This one is quite consistent with the Northern Arm and the Western Arc being related structures. On the bottom panel, the most signicant deviations from Keplerian motion discussed in the text are labeled A to D, and indicated as lled contour.

However, even though the velocity map inside is essentially that of a ow rather than that of an expanding bubble, the velocity eld is highly perturbed, and a good t with Keplerian motion must not be expected for this region. The well-known shape of the Minicavity comprises a bright spot coincident with IRS 13E on the northern end of its western edge. This bright spot is not seen in the ux map of the Northern Arm, Which contains the Minicavity. It is rst due to

283
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 13

which ionized fronts form the Minispiral, locally excited. For instance, the Bar Overlay, which velocity map is very similar to that of the Bar, may be a region belonging to the same neutral cloud as the Bar, locally excited by IRS 14W. The [He ]/[Br] line ratio is signicantly higher for these tenuous features than for the standard Northern Arm, Eastern Arm and Bar. This ratio is variable across each structure. These variations must be explained. They can basically have two reasons: rst, they can be the trace of local enrichment of the gas in helium, and second, they can be due to local enhancements of the excitation, either because of a stronger UV eld, or because of shocks. As has already been mentioned in Paumard et al. (2001), there are about 20 high mass loss stars in the region. A typical mass loss rate for stars of these spectral types is of the order of 104 M yr1 (Najarro et al. 1994). This material must reside in the central parsec for a duration similar to the time-scale of the Northern Arm: 104 yr. From these considerations, the total mass of interstellar gas in the central parsec coming from the mass loss of these stars must be around a few tens of solar masses. In the other hand, if the ionized structures are really the ionized front of neutral clouds, these clouds could have a mass similar of that of the clouds that form the Circumnuclear Disk: 103 M each (Christopher & Scoville 2003). Thus, it is unlikely that interstellar gas of stellar origin contributes signicantly to the enrichment of these clouds. However, there is a clear correlation between the projected proximity of gas to the helium stars and the [He ]/[Br] line ratio: two of the gas patches detected in both lines having a high [He ]/[Br] ratio are coincident with the helium star IRS 34W; the Bar, which is the main feature with the highest line ratio, is close to the IRS 16 helium star cluster, and contains the IRS 13E star cluster, which is made of several highmass-loss-rate stars (Maillard et al. 2003); the Tip, the feature with the highest [He ]/[Br] ratio, seems to be interacting with a star wind through the Microcavity and is on the same line of sight as the helium star IRS 9W; the dierences in the shape of the Northern Arm between the two spectral lines seem to clearly come from the geometry of the UV eld around the IRS 16 cluster. The discrepancies that are most dicult to explain are the high brightness in He , in contrast to their relative faintness in Br at the south-western parts of both the Minicavity and the Tip. However, this part of the Minicavity is rather close in projection to the AF star, which ux, if this proximity is not only in projection, could favor He emission. Finally, our results remain consistent with a well mixed interstellar material, distributed in a non-uniform UV eld, which exact value at a given point depends on the 3D localization of nearby hot stars. In addition to that, a Microcavity has been discovered at the elbow between the Eastern Arm Ribbon and Tip. It is probably a new example of interaction between stellar wind or polar jet and an ISM cloud, similar to the Minicavity. The deviation from Keplerian motion detected in the Northern Arms velocity map close to IRS 1W is also interesting, as it is probably

due to the interaction of the ow and the wind of this WR star (Tanner et al. 2003). These interactions show that the dynamics of the ows must be inuenced by the stars, as Yusef-Zadeh & Wardle (1993) suggested for the wind of the IRS 16 cluster, and the associated energy dissipation must be taken into account to nally gure out the accretion rate.

7. Conclusion
The dierent points discussed above show that the stellar and interstellar contents of the Galactic Center must be studied together to better understand them. The 2D maps of the [He ]/[Br] line ratio are in themselves interesting, but we cannot fully understand them without the complete knowledge of the 3D distribution of the interstellar material, as the line strengths are related not only to the relative abundances of the dierent elements, but also to the ionizing UV eld, which in turn depends on the distance to the ionizing sources, mainly the IRS 16 cluster, and on complex shadowing eects from the multiple ISM components. The knowledge of the radial velocity eld of the Northern Arm has allowed us to propose a kinematic model, which provides a three dimensional map of this feature. Having such maps for all of the ISM features would give us the opportunity to directly understand the shadowing eects, and to estimate the UV eld that hits these ISM features. In addition to this 3D map of the Northern Arm, we begin to gain access to the relative positions of features along the line of sight: the Eastern Bridge is closer to the observer than the Northern Arm, and the Bar is behind the Minicavity. It would then become possible to estimate the helium abundance in the dierent structures from their relative line ratios. This in turn would give a clue to the origin of these structures. This work has been performed on a eld covering most of the inner parts of the Minispiral. However, repeating the same analysis on a wider eld, containing the Minispiral to its full extent, would allow to directly check whether the Northern Arm and the Western Arc are related features. Moreover, obtaining the velocity maps of a wider eld would allow one to better constrain the parameters of the Keplerian t to the Northern Arm, and may then reveal deviations to the Keplerian model, due to momentum loss. This would be a very interesting clue to the accretion process. This scientic program requires a wide-eld spectro-imager with spectral and spatial resolutions comparable with those of BEAR.
Acknowledgements. We are grateful to Miville-Desch nes3 from the e Institut dAstrophysique Spatiale (Orsay France) for giving us the original version of the IDL spectral decomposition package and helping us in the early stages of customizing and expanding it.

References
Aitken, D. K., Smith, C. H., Moore, T. J. T., & Roche, P. F. 1998, MNRAS, 299, 743
3 Currently at the Canadian Institute for Theoretical Astrophysics (Toronto Canada).

284
14

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

Christopher, M. H. & Scoville, N. Z. 2003, in Active Galactic Nuclei: from Central Engine to Host Galaxy, meeting held in Meudon, France, July 23-27, 2002, Eds.: S. Collin, F. Combes and I. Shlosman. ASP (Astronomical Society of the Pacic), Conference Series, Vol. 290, p. 389., 389 Figer, D. F., McLean, I. S., Becklin, E. E., et al. 2000, in Proc. SPIE Vol. 4005, p. 104-112, Discoveries and Research Prospects from 8- to 10-Meter-Class Telescopes, Jacqueline Bergeron; Ed., 104112 Geballe, T. R., Krisciunas, K., Bailey, J. A., & Wade, R. 1991, ApJ, 370, L73 Genzel, R., Pichon, C., Eckart, A., Gerhard, O. E., & Ott, T. 2000, MNRAS, 317, 348 Ghez, A., Salim, S., Hornstein, S., et al. 2003, ApJ, soumis, astro-ph/0306130 Jackson, J. M., Geis, N., Genzel, R., et al. 1993, ApJ, 402, 173 Lacy, J. H., Achtermann, J. M., & Serabyn, E. 1991, ApJ, 380, L71 Lo, K. Y. & Claussen, M. J. 1983, Nature, 306, 647 Maillard, J. P. 1995, in ASP Conf. Ser. 71: IAU Colloq. 149: Tridimensional Optical Spectroscopic Methods in Astrophysics, 316 Maillard, J. P. 2000, in ASP Conf. Ser. 195: Imaging the Universe in Three Dimensions, 185 Maillard, J. P., Paumard, T., Stolovy, S. R., & Rigaut, F. 2003, A&A, submitted Menten, K. M., Reid, M. J., Eckart, A., & Genzel, R. 1997, ApJ, 475, L111 Morris, M. & Maillard, J. P. 2000, in ASP Conf. Ser. 195: Imaging the Universe in Three Dimensions, 196 Najarro, F., Hillier, D. J., Kudritzki, R. P., et al. 1994, A&A, 285, 573 Ott, T., Eckart, A., & Genzel, R. 1999, ApJ, 523, 248 Ott, T., Sch del, R., Genzel, R., et al. 2003, The Messenger, o 111, 1 Paumard, T., Maillard, J. P., Morris, M., & Rigaut, F. 2001, A&A, 366, 466 Paumard, T., Maillard, J. P., & Stolovy, S. R. 2003, in The Central 300 parsecs, Galactic Center Workshop, held in Kona, Nov. 4-8, 2002 Roberts, D. A. & Goss, W. M. 1993, ApJS, 86, 133 Scoville, N. Z., Stolovy, S. R., Rieke, M., Christopher, M., & Yusef-Zadeh, F. 2003, ApJ, submitted (astro-ph/0305350) Tanner, A., Ghez, A., Morris, M., & Becklin, E. 2003, in The Central 300 parsecs, Galactic Center Workshop, held in Kona, Nov. 4-8, 2002 Telesco, C. M., Davidson, J. A., & Werner, M. W. 1996, ApJ, 456, 541 Vollmer, B. & Duschl, W. J. 2000, New Astronomy, 4, 581 Wright, G. S., McLean, I. S., & Bland, J. 1989, in Infrared Spectroscopy in Astronomy, 425428 Yusef-Zadeh, F., Roberts, D. A., & Biretta, J. 1998, ApJ, 499, L159 Yusef-Zadeh, F. & Wardle, M. 1993, ApJ, 405, 584 Zhao, J. & Goss, W. M. 1998, ApJ, 499, L163

Appendix A: Velocity maps


We present the radial velocity maps for each identied velocity structure. Axes show osets from Sgr A (represented as a cross), in arcseconds. To improve contrast, the color scale is not the same for each map. The binocular-shaped black line shows approximately the eld boundaries.

10 5

d (arcsec)

0 5 10

20

15

10

10

15

d . cos() (arcsec)
300 150 0 150

Radial Velocity (VLSR, km . s1)

Fig. A.1. (a) Northern Arm Velocity map . The black solid line represents the constraint line (see Sect. 5.2) used for our best model. The dashed one represents another constraint line, that gave consistent results. The red contour outlines the region where two lines are detected. For these points, the bluer velocity, corresponding the the secondary layer (see text), is used. The velocity of the main layer for these points goes from 90(north-east) to 180 km s1 (south-west).

10 5

d (arcsec)

0 5 10

20

15

10

10

15

d . cos() (arcsec)
250 100 50 200

Radial Velocity (VLSR, km . s1)

Fig. A.2. (b) Bar Velocity map

285
T. Paumard et al.: Kinematic and structural analysis of the Minispiral
10 5 10 5

15

d (arcsec)

0 5 10

d (arcsec)
20 15 10 5 0 5 10 15

0 5 10

20

15

10

10

15

d . cos() (arcsec)
130 160 190 220 250 140

d . cos() (arcsec)
60 20 100

Radial Velocity (VLSR, km . s1)

Radial Velocity (VLSR, km . s1)

Fig. A.3. (c) Ribbon Velocity map

Fig. A.6. (f) Western Bridge Velocity map

10 5

10 5

d (arcsec)

d (arcsec)
20 15 10 5 0 5 10 15

0 5 10

0 5 10

20

15

10

10

15

d . cos() (arcsec)
30 70 110 150 190 220

d . cos() (arcsec)
250 280 310 340

Radial Velocity (VLSR, km . s1)

Radial Velocity (VLSR, km . s1)

Fig. A.4. (d) Eastern Bridge Velocity map

Fig. A.7. (g) Tip Velocity map

10 5

10 5

d (arcsec)

0 5 10

d (arcsec)
20 15 10 5 0 5 10 15

0 5 10

20

15

10

10

15

d . cos() (arcsec)
40 10 20 50 80 10

d . cos() (arcsec)
30 50 70

Radial Velocity (VLSR, km . s1)

Radial Velocity (VLSR, km . s1)

Fig. A.5. (e) Western Arc Velocity map

Fig. A.8. (h) Northern Arm Chunk Velocity map

286
16
10 5

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL


T. Paumard et al.: Kinematic and structural analysis of the Minispiral

10 5

d (arcsec)

0 5 10

d (arcsec)
20 15 10 5 0 5 10 15

0 5 10

d . cos() (arcsec)
260 180 100 20

20

15

10

10

15

d . cos() (arcsec)
Radial Velocity (VLSR, km . s1)
Fig. B.2. (b) Bar ux map. A few contours of the ux map of the Northern Arm are given (dotted lines), showing that the Minicavity corresponds to a low level region of the Bar (Sect. 4.4).

Fig. A.9. (i) Bar Overlay Velocity map

Appendix B: Flux maps


Flux map (peak intensity width of the line) for each structure. Black outlines give the full extent of the detected structure, and eld boundaries. Axes are osets from Sgr A , which position is marked by a cross.

10 5

d (arcsec)

0 5 10

10 5

d (arcsec)

0 5 10 10 20 15 10 5 0 5 10 15 5

20

15

10

10

15

d . cos() (arcsec)
Fig. B.3. (c) Ribbon ux map

Fig. B.1. (a) Northern Arm ux map. The outline of the Eastern Bridge is given (dashed line), showing that this structure is superimposed on a weak region of the Northern Arm (Sect. 4.4).

d (arcsec)

d . cos() (arcsec)

0 5 10

20

15

10

10

15

d . cos() (arcsec)
Fig. B.4. (d) Eastern Bridge ux map

287
T. Paumard et al.: Kinematic and structural analysis of the Minispiral 17

10 5 0 5 10

10 5

d (arcsec)
20 15 10 5 0 5 10 15

d (arcsec)

0 5 10

20

15

10

10

15

d . cos() (arcsec)
Fig. B.5. (e) Western Arc ux map

d . cos() (arcsec)
Fig. B.8. (h) Northern Arm Chunk ux map

10 5

10 5

d (arcsec)

0 5 10

d (arcsec)
20 15 10 5 0 5 10 15

0 5 10

20

15

10

10

15

d . cos() (arcsec)
Fig. B.6. (f) Western Bridge ux map

d . cos() (arcsec)
Fig. B.9. (i) Bar Overlay ux map

10 5

d (arcsec)

0 5 10

20

15

10

10

15

d . cos() (arcsec)
Fig. B.7. (g) Tip ux map

288

ANNEXE E. ARTICLE KINEMATIC ANALYSIS OF THE MINISPIRAL

Annexe F Article Structural analysis of the Minispiral from high-resolution Br data


Contribution pour la confrence Galactic Center workshop 2002, relative au travail sur la Minispirale.

289

290

ANNEXE F. ARTICLE STRUCTURAL ANALYSIS OF THE MINISPIRAL

291

Astron. Nachr./AN 324, No. S1, 1 8 (2003) / DOI theDOIprex/theDOIsufx

Thibaut Paumard1 , Jean-Pierre Maillard1 , and Mark Morris2


1 2

Institut dastrophysique de Paris (CNRS), 98b Bd. Arago, 75014 Paris, France University of California, Los Angeles, Div. of Astronomy, Dept. of Physics and Astronomy, Los Angeles, CA 90095-1562, USA

Received 15 November 2002, revised 30 November 2002, accepted 2 December 2002 Published online 3 December 2002 Key words dynamics, ionized gas, Sgr A West , Galaxy: Center, infrared, spectro-imaging, FTS PACS 04A25 Integral eld spectroscopy of a roughly 40 40 region about the Galactic Center was obtained at 2.16 (Br ) using BEAR, an imaging Fourier Transform Spectrometer, at a spectral resolution of 21.3 , and a spatial resolution of 0.5 . The analysis of the data was focused on the kinematics of the gas ows concentrated in the neighborhood of SgrA , traditionally called the Minispiral. From the decomposition into several velocity components (up to four) of the line prole extracted at each point of the eld, velocity features were identied. Nine distinguishable structures are described: the standard Northern Arm, Eastern Arm, Bar, Western Arc, as well as ve additional moving patches of gas. From this analysis, the Northern Arm appears not limited, as usually thought, to the bright north-south lane seen on intensity images, but consists instead of a continuous, weakly-emitting, triangular-shaped surface having a bright western rim, and narrowed at its forward apex in the vicinity of SgrA where a strong velocity gradient is observed. The gravitational eld of the central Black Hole can account for both the strong acceleration in this region and the tidal compression of the forward tip of the Northern Arm. Keplerian orbits can be tted to the velocity eld of the bright lane, which can be interpreted as formed by the bending of the western edge of the owing surface. These results raise questions regarding the formation of the Sgr A West gas structures.
      

1 Introduction
Within the inner 2 pc of the Galactic Center (GC) lies the Sgr A West region, dominated by ionized gas which has been detected in the infrared and at radio wavelengths because of high obscuration along the line of sight. The gas distribution has been observed in infrared and radio emission lines, as well as in radio continuum (Lacy et al. 1991, Lo & Claussen 1983, Roberts & Goss 1993, Zhao & Goss 1988), and proper motion of the bright blobs have been derived (Zhao & Goss 1988). All these data show that the ionized gas in the inner few parsecs of Galactic Center is organized into a spiral-like feature (Fig. 1) with a number of arms, which led to the name Minispiral for the entire pattern. The various features give a spiral appearance primarily because of the way they are superposed on each other. However, a new analysis of Lacys data was conducted by Vollmer & Duschl (2000) to re-examine the kinematic structure of the ionized gas. Using a three-dimensional representation, they conrm the standard features, but with a more complex structure, including two features for the Eastern Arm, a vertical nger of high density, a large ribbon extending to the east of SgrA , and two distinctly different components in the Bar. Data in different lines, acquired with better spectral and spatial resolution, warrant an independent kinematic analysis. In the present paper, the gas content in the inner region of the GC is presented and analyzed from high spectral resolution (21.3 ) data cubes on the Br line, obtained with BEAR, an imaging FTS
c 2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 0004-6337/03/S101-0001 $ 17.50+.50/0

Structural analysis of the Minispiral from high-resolution Br data

'

% $ " &#!

292

ANNEXE F. ARTICLE STRUCTURAL ANALYSIS OF THE MINISPIRAL

T. Paumard, J.P. Maillard, and M. Morris: Structural analysis of the Minispiral

Fig. 1 Br intensity integrated between and toward the two mosaicked elds of Sgr A West observed with BEAR. The standard bright features, Northern and Eastern Arms, Bar, and the mini-cavity, are indicated. Also, a few emission line stars show up as bright points in the image. (Morris & Maillard 2000)

(Maillard 1995, 2000) at the f/35 infrared focus of the 3.6-m CFH Telescope; the same work on He I data is ongoing. A preliminary analysis of these data have been presented in Morris & Maillard (2000). In Paumard et al. (2001, hereafter Paper I), where the helium stars were studied, the He I ISM emission was at a seeingidentied and described for the rst time. The Br data cover a eld of view of limited resolution of 0.6 . The HST-NICMOS Pa data (Scoville et al. 2003), used in Paumard et al. (2003), are used here for high-resolution morphological considerations. A multi-component line tting procedure applied to the emission-line proles in each point of the eld is described in Sect. 2. From this decomposition, the identication of dened gas structures constituting the whole Sgr A West ionized region is presented in Sect. 3. Attempts to adjust Keplerian orbits to the owing gas are presented. This is followed in Sect. 4 by a discussion of the implication of these identications on the formation and the lifetime of the inner ionized gas.

2 Structure identication
In each point of the eld (Fig. 1) the Br emission prole appears complex, clearly showing that along each line of sight, several ISM clouds or ows are present. So the study of these ISM features required the development of a multi-component line tting procedure able to work on 3D data. From a coarse examination of the datacube the tting by a maximum of four components seemed adequate. MivilleDesch ne (private communication) provided us with such software, that we adapted and developed. By e comparison of the velocity components from one line of sight to the next, and by assuming continuity, it is possible to perform a reconstruction of the large-scale velocity structures. In the end, it might be possible to conclude whether these structures are independent, or continuous spiraling ows. Thus, the process is split into two main parts: rst, the line prole decomposition of all the points of the eld, and then the structure identication. The natural line shape of a single velocity component of the emission line from the ISM has been assumed to be Gaussian, dened by three parameters (central velocity, amplitude, and line width). The tting procedure then adjusts four independent Gaussian lines, convolved with the instrumental PSF which is a sine cardinal to each spectrum. Hence the tting function depends upon twelve independent parameters for each point of the eld. The consecutive steps are the following: Preparation: These twelve parameters are rst manually determined for a few starting points, evenly distributed over the eld, chosen to be representative of the most obvious features, and located on spots where four components are clearly present.

" !&$ "" "'% #

    !

""

293

Astron. Nachr./AN 324, No. S1 (2003)

Step 1: From the starting points, a rst procedure attempts to t a four-component line shape function to each spectrum. For each new spectrum, the initial guess is determined from the results found for the neighboring points. The spectra are studied sequentially in parallel spiral-mode scannings arround each starting point. Except for the initial guess, the tting of a spectrum is independent of all the others. Step 2: For the brightest point of the eld, the neighbors are examined, and searched for a component such that the velocity gradient between the point of interest and this neighbor is less than a certain amount, which is set by the user at runtime. The procedure is iterative, and once a few neighbors have been selected into a structure, their neighbors are in turn examined for possible selection. The procedure stops when every component of every point of the eld has been assigned to exactly one spatial structure. This procedure allows only one component of a given point to be selected into a given structure. Structures so warped that they overlap themselves spatially, thus causing two velocity components on the same line of sight, cannot be directly detected as such: they articially split into several structures. Step 3: This procedure requires that the detected structures be manually inspected. The user has then the possibility to add some more common sense heuristics in the structure identication, a little difcult to implement but easy to apply manually. Several problems can occur: during step 1, the tting procedure might t only one component where two blended components are indeed more appropriate,

during step 2, the procedure can falsely cross-connect two structures, i.e. reconstruct two structures, each one being made of one part of one physical structure, and one part of the other one.

Step 4: Then, these manually corrected results are used to perform a second t on each point of the eld; at this point, 2D information is entirely included in the initial guess provided to the tting procedure. Iteration: Steps 2, 3 and 4 must be iterated a number of times, until a stable set of plausible structures is reached. Plausible structures means only that the structures are more extended than the spatial resolution, and brighter than the detection limit of the instrument.

3 Results
3.1 General description of the results The analysis described above leads to a vision of the Minispiral more complex than usually thought, one which is consistent with, but more detailed than the vision proposed by Vollmer & Duschl (2000). After a careful examination we identify 9 components of various sizes (Fig. 2). Two types of velocity map appear, some with an overall velocity gradient, others without any large-scale velocity gradient. The deviation from mean motion, dened as the local difference between the velocity measured at one point and the mean value for the neighboring points, and divided by the uncertainty, ranges from roughly one tenth to ten for all the features, which means that every velocity structure shows signicant (over 3 ) local features. The areal size of the structures (Table 1), expressed in terms of solid angle covered on the sky, ranges from 17 arcsec to 300 arcsec for the part of the Northern Arm that is visible in the BEAR eld of view. The surface area of each structure must be considered as a lower limit because BEAR may not detect the weakest parts nor parts where blending with a brighter structure in the spectral domain is possible, and because the eld of view does not cover the entire Minispiral. 3.2 Morphology of the ionized gas in Sgr A West A brief description follows for each identied velocity structure, whose velocity maps are given in Fig. 2. Table 1 gives the surface coverage on the sky (within the BEAR eld of view), and the maximum and minimum velocity within the given structure.

294

ANNEXE F. ARTICLE STRUCTURAL ANALYSIS OF THE MINISPIRAL

T. Paumard, J.P. Maillard, and M. Morris: Structural analysis of the Minispiral

a)
f

b)
i h

c) e)

Fig. 2 Velocity maps for the nine detected structures (see text for precise identication). Coordinates are offsets from SgrA , in arcseconds.

d) f) S (arcsec ) 300.8 173.1 103.8 83.5 58.7 40.7 25.8 23.1 16.9

g) h) i)

a) Northern Arm: Contrary to its standard description, the Northern Arm is not seen here as a bright N-S lane, but as an extended, triangular surface. One edge of this triangle is the bright rim generally noticed, but it extends over to the Eastern Arm. The third edge of the triangle is the edge of the eld, so viewing this feature on a larger eld may yield a slightly different description. Its kinematics will be thoroughly described in the next section. b) Bar: The Bar is the most complex region, where at least three components are superimposed. The most important feature is very extended, from the Eastern Arm (c) to the Western Arc (e), very straight and shows a smooth overall velocity gradient. Vollmer & Duschl (2000) mention two complementary components of the Bar, which they call Bar 1 and Bar 2, though their description is not sufcient to determine precisely the positions of these two components. We see two additional features, which we propose to call Western Bridge (f) and Bar Overlay (i). Parts of the Bar are also superimposed on almost every other structure, including the Eastern Arm (c), the Tip (g), the Eastern Bridge (d) and the Northern Arm (a). c) Eastern Arm: The region is split into two parts: the Arm itself and a Tip (g). The velocity gradient of the Arm is directed along the minor axis of the structure, not along its major axis as expected for a ow. d) Eastern Bridge: A structure of medium size extends from the Eastern Arm to the bright rim of the Northern Arm. It does not show any large-scale velocity gradient, and its shape does not show any principal axis that would indicate a ow. It is superimposed on the faint regions of the Northern Arm, and partly superimposed on the Eastern Arm, the Bar and the Tip. The Pa map (Fig.5) shows that this feature may extend outside our eld-of-view into a elongated feature parallel to the Eastern Arm.


ID a b c d e f g h i

Feature name Northern Arm Bar Eastern Arm Eastern Bridge Western Arc Western Bridge Tip Northern Arm Chunk Bar Overlay

S (pix) 2414 1389 833 670 471 327 207 185 136

Vmin -286.9 -211 132.9 34.9 -37.1 -121.1 222.9 14.9 -267.0

Vmax 188.9 196.9 242.9 182.9 74.9 100.9 339.0 74.9 -7.1

Table 1 Feature identications, with surface areas (pixels and square arcseconds), and minimum and maximum radial veloc). ities (

295

Astron. Nachr./AN 324, No. S1 (2003)

Fig. 3 Microcavity feature in the region of IRS 9W, which would be approximately located in the upper right corner. The center of this eld is located 5.4 east and 8.1 south of SgrA .

e) Western Arc: The Western Arc lays just at the edge of the eld, so we have access only to its innermost part. It is seen as a rather simple feature, with large scale velocity gradient. f) Western Bridge: The Western Bridge is a tenuous, elongated feature oriented east-west and extending from the Bar to the Western Arc. g) Tip: The Tip is, in projection, a very concentrated and relatively small object with the most redward ). The Tip has already been noticed by Vollmer & Duschl (2000) velocity in the region ( 300 only on a morphological basis, as a nger-like feature of the Eastern Arm in their three dimensional data. Here, we see that the Eastern Arm and the Tip are two distinct features, superimposed on the line of sight, thus we do not adopt the representation-dependent denomination Finger. At the elbow between the Eastern Arm and the Tip, in the IRS 9W region, is a bubble-like feature, or a Microcavity (radius 1 ), with a rather bright rim (Fig. 3), which appears at a specic velocity (250 ). h) Northern Arm Chunk: A small tenuous structure is seen superimposed on the Northern Arm, a few arcseconds north of IRS 7. It lays at the edge of our eld, so it could extend further out; however the Pa image shows a small, horizontal bar at his location, crossing the bright rim of the Northern Arm, and that does not seem to be much extended. i) Bar Overlay: The Bar Overlay looks like a small cloud that is superimposed upon the western region of the Bar and that shows a velocity gradient similar to the one of the main Bar at the same location, . It may indicate that these two features are closely related. They could, with an offset of -40 for example, be the two faces of a single cloud. 3.3 Keplerian orbit tting The velocity maps show a view of the features very different from the usual ux maps which, by themselves, can be misleading. For instance the morphology of the Northern Arm with its typical bright rim may lead one to think of it as the true path for most of the material. On the other hand, the velocity map shows no peculiar feature at the location of the rim. This is particularly intriguing for the location where it bends abruptly, just a few arcsecond north of IRS 1 and east of IRS 7 (Fig. 1). Thus we are led to the idea that the kinematics of the Northern Arm should be studied independently of its intensity distribution. As a rst attempt at exploiting the information contained in the new tools that are the velocity maps, we tried to analyse the Northern Arm as a Keplerian system. For a rst, simple approach, we created a dedicated IDL graphical package called GuiMapOverlay (Fig. 4). With this tool, the user can easily adjust one Keplerian orbit over a velocity map. A Keplerian orbit in 3D is dened by ve orbital parameters: the eccentricity, two angles dening the orientation of the orbital plane, the periapse (distance of closest approach to the center of motion), and a third angle dening the position of the periapse. Once the user is almost satised with the orbital parameters he has found by trial and error, an automatic tting procedure can be called. It is possible to x parameters (at least two must remain free), and the orbit can be forced to go through a selected constraint point by tying the periapse to the other parameters. After a few experiments with this tool, we are led to some general conclusions:

  

 

  



296

ANNEXE F. ARTICLE STRUCTURAL ANALYSIS OF THE MINISPIRAL

T. Paumard, J.P. Maillard, and M. Morris: Structural analysis of the Minispiral

Fig. 4 GuiMapOverlay: given a set of orbital parameters, this tool draws the path of the corresponding Keplerian orbit on a velocity map, extracts the observed velocities along this path, and plots both the observed and computed velocity curves. The is also computed and shown.

a good agreement can be found between observed and calculated velocities, except in the region of the Minicavity; this model alone is not sufcient to decide whether the orbits are bound or not, or whether the data are compatible with elliptic, parabolic, or hyperbolic motion.

The second point mentioned above is not satisfactory, as one of the most interesting questions is to decide whether the gas is bound. Nevertheless, the rst remark encouraged us to persevere in the direction of Keplerian modeling, so we attempted to model the Northern Arm with several orbits instead of only one. Each orbit is bound to pass through a different point, by tying its periapse to the other four parameters, and all these constraint points are aligned across the gas lane, close to the line of zero radial velocity, thus ensuring that they are indeed on separate trajectories. To ensure a smooth model we are interested only in the global motion the four functions that map each constraint point to one of the parameters, have been chosen to be described as spline functions, dened by their value at four points. Thus, having four functions (one for each of the orbital parameters), each of them being dened by four values, the model is dependent on sixteen parameters. We designed a tting procedure to adjust this model based on the observed velocity map by minimizing the reduced . In order to study a certain hypothesis, it is possible to either x some parameters, or to force them to have the same value for each orbit. This way, for example, it is possible to really check whether the observed velocity map is consistent with coplanar orbits or with uniform eccentricity. To avoid studying only local minima in the parameter space, it is also important to use several different initial guesses. A tool quite similar to GuiMapOverlay has been designed to easily study whether the data are consistent with an homothetic1 set of orbits, which is the simpler model. 3.3.1 Results With the homothetic set of orbits hypothesis, the eccentricity still cannot be well constrained. Bound orbits seem to be preferred, but the agreement is as good with circular orbits and very elliptic orbits, close to parabolic. The residuals map has always the same shape: the observed velocities are always smaller than the computed ones along the inner edge of the bundle of orbits, and higher along the outer edge. The . A few of these homothetic models have been chosen as agreement is always poor, with initial guesses for other adjustments, with released constraints. It is rst interesting to check the coplanar hypothesis, by keeping uniform only the two parameters that dene the orbital plane, and the uniform
1 Two orbits are said to be homothetic when they are identical except for their scale, i.e. when they share the same orbital

parameters, except the periapse.

$" %#!    

 

297

Astron. Nachr./AN 324, No. S1 (2003)

eccentricity hypothesis. The agreement is much better when making either the eccentricity or the orbital plane free, but it is still better making both free. Even with the most general situation, the parameters are still not constrained enough to decide whether the orbits are all bound or not, to extrapolate the model outside the eld of view, nor even to derive reliably the direction of proper motion. However the models share a few characteristics that we judge to be robust because of their repeatability: 1. the orbital planes are close to that of the CND; 2. the orbits are not quite coplanar; the two angles that dene the orbital plane vary over a 10 range; 3. the eccentricity varies from one orbit to another, beeing close to parabolic or above for the innermost orbits, and closer to circular (below 0.5) for the outermost. The variations of the orbital parameters induce a particular shape for the Northern Arm (Fig. 5): for all the non-coplanar models, the Northern Arm looks like a warped surface, and this warp induces a crowding of orbits that closely follows the bright rim of the structure. That suggests that the Northern Arm is either a warped disk, or the ionized surface of a neutral cloud. The bright rim itself is not due only to the stronger UV eld and a real local enhancement of the density, but also to an enhancement of the column-density due to the warp. An interesting point is that, in some models, no orbit follows the bright rim, which emphasizes that it is really important to consider the dynamics independently from the morphology of the Northern Arm. Another characteristic present in all the models is that the period of the orbits ranges from a few years to a few years, which implies that the Northern Arm would have a completely different shape in a few years, and cannot be much older than that timescale. Since the agreement in radial velocity is now rather good, it makes sense to look at the deviations from global motion by looking at the extended features on the residual velocity map (Fig. 5): A) the ow shows a rather signicant deviation in the region just southwest of the embedded star, IRS 1W; this perturbation could be due to the interaction with this stars wind; B) the region of this model closest to the Minicavity is perturbed; C) another deviation is seen at the precise location where the bright rim bends a lot, just east of IRS 7E2; D) nally, an elongated feature is seen on the fainter rim coming from IRS 1W towards the northeast.

4 Discussion
The presence of at least three isolated gas patches (the Western Bridge, the Northern Arm Chunk and the Bar Overlay, but also possibly the Eastern Bridge, which may or may not extend ouside the eld) in addition to the standard large ows has been demonstrated. In addition to that, the Microcavity at the elbow between the Eastern Arm and the Tip is a new example of an interaction between an ISM feature and a stellar wind, similar to the Minicavity, as is the deviation from Keplerian motion detected close to IRS 1W. In this context, it makes sense to ask what is the inuence of the large number of mass losing stars present in the central parsec? These massive, hot stars of the central cluster, named helium stars from their strong 2.06 He I emission line, presumably LBV-type and WR stars, being particularly concentrated in two clusters, IRS 16 (Krabbe et al. 1991) and IRS 13E (Maillard et al. 2003), must be a major source of helium in their environment. Therefore the following question arises: what happens to this helium enriched material? Could it form or enrich the gas patches that we see? Comparing the helium and hydrogen distribution in the central parsecs, and their abundance in the different structures and in the CND is guaranteed to help us better understand the origin of the different ISM structures. The geometry of the Northern Arm has been studied from its velocity map, leading to the conclusion that it may not be a planar structure, but rather a three-dimensional structure. Fig. 5a is quite compatible with the Northern Arm indeed being the ionized surface of a neutral cloud, which could come from the CND, in accordance with the standard formation scenario.
 

298

ANNEXE F. ARTICLE STRUCTURAL ANALYSIS OF THE MINISPIRAL

T. Paumard, J.P. Maillard, and M. Morris: Structural analysis of the Minispiral

Fig. 5 On this Pa map (Scoville et al. 2002), one of the Keplerian models is overplotted (left panel). This one is quite consistent with the Northern Arm and the Western Arc being related structures. On the right panel, the most signicant deviations from Keplerian motion discussed in the text are labeled A to D, and indicated as lled contour.

References
Krabbe, A., Genzel, R., Drapatz, S., & Rotaciuc, V. 1991, ApJ, 382, L19 Lacy, J.H., Achtermann, J.M., & Serabyn, E. 1991, ApJ, 380, L71 Lo, K.Y., & Clausen, M.J. 1983, Nature, 306, 647 Maillard, J.P. 1995, In: Tridimensional Optical Spectroscopic Methods in Astrophysics, IAU Col. 149, G. Comte & M. Marcelin (eds), ASP Conf. Series, 71, 316 Maillard, J.P. 2000, in Imaging the Universe in Three Dimensions, E. van Breughel & J. Bland-Hawthorn (eds), ASP Conf. Serie 195, 185 Maillard, J.P., Paumard, T., Stolovy, S.R., & Rigaut, F. 2003, these proceedings Morris, M. & Maillard, J.P., 2000, in Imaging the Universe in Three Dimensions, E. van Breughel & J. BlandHawthorn (eds), ASP Conf. Serie 195, 196 Paumard, T., Maillard, J.P., Morris, M., & Rigaut, F. 2001, A&A, 366, 466 (Paper I) Paumard, T., Maillard, J.P., Stolovy, S.R., & Rigaut, F. 2003, these proceedings Roberts, D.A. & Goss, W. M. 1993, ApJS, 86, 133 Scoville, N.Z., Stolovy, S.R., Rieke, M., Christopher, M., & Yusef-Zadeh, F., 2002, submitted to ApJ Vollmer, B., & Duschl, W.J. 2000, New Astronomy, 4, 581 Yusef-Zadeh, F., Roberts, D.A., & Biretta, J. 1998, ApJ, 499, L159 Yusef-Zadeh, F., Stolovy, S. R., Burton, M. Wardle, M., & Ashley, M. C. B. 2001, ApJ, 560, 749 Zhao, J.H. & Goss, W.M. 1988, ApJ, 499, L163

Annexe G Article The star-forming region S106 revealed by BEAR spectro-imagery

299

300

ANNEXE G. ARTICLE THE STAR-FORMING REGION S106

You might also like