Proc. Nat. Acad. Sci. USA Vol. 71, No. 11, pp. 4496-4499, November 1974

Chemical Composition of Nebulosities in the Magellanic Clouds (gaseous nebulae/galaxies) L. H. ALLER*t, S. J. CZYZAKtt, C. D. KEYES*, AND G. BOESHAARt t Department of Astronomy, The Ohio State University, Columbus, Ohio; and *Department of Astronomy, University of California, Los Angeles, Calif. 90024 Contributed by L. H. Aller, August 14, 1974

ABSTRACT From photoelectric spectrophotometric quate accuracy, application of the relevant theory to observa- data secured at Cerro Tololo Interamerican Observatory tions suffers difficulties and frustrations. we have attempted to derive electron densities and tem- spectral line emission per cm3 peratures, ionic concentrations, and chemical abundances The equations express the of He, C, N, 0, Ne, S. and Ar in nebulosities in the Magel- as a function of ionic concentration, Nf, electron density, AT. lanic Clouds. Although 10 distinct nebulosities were and gas kinetic electron temperature, TE. We actually observe observed in the Small Cloud and 20 such objects in the the emission from a radiating column of gas of many parsecs Large Cloud, the most detailed observations were secured depth and cross-section within which both and can only for the brighter objects. Results for 30 Doradus are in T. N, harmony with those published previously and recent work fluctuate considerably. Peimbert (9) has calculated the in- by Peimbert and Torres-Peimbert. Nitrogen and heavier fluence of a mean square fractional temperature fluctuation elements appear to be less abundant in the Small Cloud ((T - To)/To)2 upon the emmissivity of a spectral line in a than in the Large Cloud, in accordance with the conclu- medium of average temperature To. sions of Dufour. A comparison with the Orion nebula may be observed only in suggests He, N, Ne, 0, and S may all be less abundant in Some elements such as nitrogen the Magellanic Clouds, although adequate evaluations the neutral or singly ionized stages. In a typical ionized will require construction of detailed models. For example, hydrogen (HII) region, most of the nitrogen exists as un- if we postulate that the [NIl], 10111, and [S111 radiations observable N++. Seaton (10) and Peimbert and Costero (11) originate primarily in regions with electron temperatures have proposed recipes for allowing approximately for the near 8000'K, while the [0111], [NelIl], [ArIbl, and H radi- ionization ations are produced primarily in regions with TE = 10,000° distribution of various atoms among different K, the derived chemical abundances in the clouds are stages by considering, e.g., n(O+)/n(O++), n(He++)/n(He+), enhanced. etc. ratios. We used their procedures only because there now way of calculating the nebular radiation the exists no satisfactory A basic assumption that underlies methods of establishing field and ionization equilibrium of elements such as nitrogen. cosmological distance scale is that the chemical composition Electron densities can be estimated from certain forbidden of stars and nebulosities in distant galaxies is essentially the line doublet ratios observed in ions with a p3 configuration same as in our own stellar system, exhibiting the same pre- (12-15); the important ratios are 3726/3729 [OIl] and 6717/ ponderance of light elements and possible spreads of metal-to- 6731 [SII]. Whenever possible, we calculated T, from the hydrogen ratios from one stellar population type to another. ratio. com- 4363/5007 [OIII] Three direct methods of assessing possible chemical For p2 and p4 ions, the ionic concentration N(ion) with position differences are immediately apparent: (i) quantita- respect to ionized hydrogen N(H+) is given by: tive spectral analyses of the brightest stars; (ii) studies of of composite spectral features and energy distributions N(ion) = a,[1 + a2x]tl/2Eo4210blt I(ne) [1] integrated light of a stellar system; and (iii) analysis of the spectra of emission nebulosities. and transition Inspired by Przybylskis ((1-3) pioneering efforts, several where a, and a2 depend on collision strength workers (4-6) have measured concentrations of abundant probabilities. elements in a number of Magellanic Cloud supergiants. Since x = 10-4NT/ Vt, t = 10-4T,, b = 0.540XIj [2] of these stars teeter on the brink of instability, atmospheres the or the usual assumptions of hydrostatic equilibrium, stratifica- where x1j is the excitation potential of (1D2) ('Se) level, and local thermodynamic equi- respectively. The emission per unit volume in HO3 is E(HO3) = tion in plane parallel layers, been calculated are to question. Spectra of integrated starlight N(H+)N10-25 E04,2. where E04,9 has by librium open expres- can give only crude estimates of metal/II ratios. Spectra of Clarke (16)§ and by Brocklehurst (17). Appropriate emission nebulosities can be studied even in distant galaxies, sions or tables (12-15) exist for p3 ions. and certainly abundance ratios can be reliably established, METHODS AND RESULTS as shown, e.g., by Peimbert and Spinrad (7, 8). Although physical processes occurring in thermally excited Optical and Radio-frequency Studies of Emission Nebulosities gaseous nebulae appear to be well understood, and required in the Magellanic Clouds. Finding charts, positions, and de- atomic parameters have been calculated to seemingly ade- scriptions of emission nebulosities, and also surface bright- & t Guest investigators at Cerro Tololo Interamerican Observatory, § Values are quoted by Aller and Liller (1968) in Table 6; Stars operated by the Association of Universities for Research in Stellar Systems 7, 483-574; Nebulae & Interstellar Matter, ed. Astronomy under contract with the National Science Foundation. Middlehurst, B. & Aller, L. H. (Univ. of Chicago Press, Chicago). 4496 Downloaded by guest on September 23, 2021 Proc. Nat. Acad. Sci. USA 71 (1974) Composition of Nebulosities in Magellanic Clouds 4497 TABLE 1. Adopted line intensities for 30 Doradus* TABLE 2. Derived t and x parameters for 30 Doradus* X(k) Identity I X(A) Identity I Derived (assumed) 3727 [OII] 93 6300 [0I] 5.9 parameter 3835 H8 5.5 6312 [SIII] 2.2 Corrected 3868 [NeIII] 25.0 6548 [NII] 8.4 Diagnostic ratio I ratio x t Refs. 3889 H,He 14.0 6563 Ha 405 4340 Hly 42.3 6584 [NII] 20.5 3729/26 (0+) 0.94 0.074 12, 36 4363 [GIll] 2.8 6678 HeI 9.6 6717/31 (S+) 1.23 0.09 (0.8) 13 4861 H# 100 6717 [SII] 14.2 4068/6725 (S+) 0.045 0.09 (0.8) 13, 36 4959 [0111] 164 6730 [SII] 11.5 3727/7330 (0+) 33.4 (0.074) 1.02 12 5007 [GIll] 499 7065 HeI 4.6 5007/4363 (O++) 125 (0.074) 1.04 40 5755 [NII] 0.25 7135 [ArIII] 22.7 5755/6548,84 (N+) 0.010 (0.074) 1.0 38, 39t 5876 HeI 13.8 7330 [OII] 6.0 * See Eq. (2) for definitions. Each ratio depends on t and x. * These line intensities have not been corrected for inter- The insensitive one is indicated by parentheses. For example, stellar extinction. 3729/3726 is sensitive to NE at relevant densities but insensitive to T,; the reverse is true for the 5007/4363 ratio. t A-values for [NII] are from ref. 38. The collision strengths nesses in Ha + [NII], have been given by Henize (18), and by are from ref. 39. Doherty et al. (19), respectively. Dickel et al. (20) published a photoelectric spectrophotometric survey of brighter emission a two-temperature model. We examine here two idealized nebulosities, and assigned excitation classes (21). Later, models: (I) uniform temperature, t = 1.03, x = 0.074 for all Dickel (22) measured isophotic contours and determined ions and (II) t = 0.95 for zones of O++, Ne++, S+*+, Ar++, mean electron densities and masses for HII regions of the Large Cloud. Ar+3, He+, x = 0.074; and t = 0.8 for N+, 0+, S+, x = 0.074. It is assumed that most of the hydrogen emission comes from From data secured at 73 cm in radio-frequency both clouds, the t = 0.95 zone. Mills and Aller (23) found smaller root mean square N, values For model I the values-of log N(ion)/N(H+) are: -5.63 than indicated by other studies. Filamentary structure can (N+), -5.53 (O°), -4.33 (O+), -3.87 (0++), -4.53 (Ne++), explain these differences. Extensive radio-frequency studies - 5.63 (S+), -5.46 (S++), and -5.93 (Ar++). Higher values in the Large Cloud were made by Mathewson and Healey (24) would be obtained for model II or by following Peimbert's at 73 and 21 cm, by J. N. Clarke (25), and by McGee and his procedure, for which one must know the root mean square associates (26-28) at 6 and 11 cm. temperature fluctuation. The Observations. No scanner was available in 1970 when Uncertainties in t (and to some extent in x), as well as this program was initiated. We obtained slit spectrograms difficulties in allowing for unobserved ionization stages, all with the 0.91-m and 1.5-im Cerro Tololo telescopes at carefully introduce inaccuracies. We used the same correction factor selected positions in each nebulosity. Unfavorable weather for n(S+) + n(S++) as for n(N+) to get the sulfur concentra- made it impossible to obtain the necessary calibrations by tion (11). Uncertainties are large, not only for S, but par- scanner measurements of individual lines and line pairs in ticularly for Ar. November 1972, so the program had to be carried out a year The derived logarithmic elemental abundances from model later. Although the present discussion is based primarily on I and model II, denoted by parentheses, are: He -1.09 photoelectric measurements, photographic spectral data (-1.10); N -5.05 (-5.01); 0 -3.72 (-3.48); Ne -4.40 supplied supplementary information, sometimes significant (-4.13); S -4.77 (-4.82), Ar -5.9 (-5.7). in assessing the importance of filamentary structures. Details The low helium abundance in 30 Doradus is of particular will be published elsewhere. interest. Faulkner and Aller (35) published a value, n(He+)/ N(H+) = 0.082 ± 0.008, that has been confirmed by Peim- 30 Doradus and NGC 1936 = Henize 44C. Many studies bert and Torres-Peimbert (33) on the basis of their recent have been made of the structure (29, 30), density, (31), and excellent observations. Apparently, this low ratio cannot be spectrum (32-36) of the great 30 Doradus nebula, both at explained by incomplete ionization of helium (33). The abun- optical and radio frequencies (26-28, 37). For the intensively dance seems definitely lower than in galactic nebulae such as observed bright arc, our I(Ha)/(Hj3) ratio appears larger Orion (11) ort- Carinae (35). than that reported by Mathis (34), although our data agree We have observed the spectra of several regions of Henize well with some previous photoelectric (35) and photographic 44, viz H44B = NGC 1935, H44C = NGC 1936, and 44D spectrophotometry (36). By assuming the same color depen- both photographically and photoelectrically. Table 3 gives a dence for interstellar extinction in the Large Cloud as in our synopsis of line intensity measurements and ionic concentra- own galaxy, we find an attenuation of 1 magnitude at HB from tions for the high excitation region in H44C which showed the Balmer decrement. Except for [OI], the then corrected X4686 HeII, and X4740 [ArIV]. Dickel (22) measured isophotic intensities agree fairly well with those of Peimbert and Torres- contours for this H44 complex, which Bok (42) identified as a Peimbert (33) for the same region (their 30 Dor II). region of star formation. McGee and Newton (28) suggested Table 1 gives intensities of lines observed photoelectrically. H44 contains a non-thermal source. A 2 radio-frequency greater Table lists the plasma diagnostic ratios used. We adopted range of excitation occurs here than in most diffuse nebulosi- t = 1.03, x = 0.074 (N. = 750 electrons/cm8). Possibly the ties. [SII] lines are produced in cooler regions (t about 0.8), or the discrepancies may arise from observational and theoretical Derived Composition of the Magellanic Clouds. For model I uncertainties. A rediscussion (41) of earlier data (36) used of the Large Megellanic Cloud we computed t from the Downloaded by guest on September 23, 2021 4498 Astronomy: Aller et al. Proc. Nat. Acad. Sci. USA 71 (1974) TABLE 3. Synopsis of line intensity measurements in Henize TABLE 4. Logarithms of elemental abundances 44C (NGC 1936)* log N(H) = 12.00

(1) (2) (3) (4) (1) (2) (3) (4) Small Large Megallanic 3727 O+ 2.28 -4.27 5007 O++ 2.80 -3.72 Megellanic Cloud 3868 Ne++ 1.82 -4.33 5876 He(+) 1.31 -0.83: Cloud I II Orion 4026 He(+) 0.32 -1.04 6300 O+ 0.96 -5.29 4471 He(+) 0.7: -0.99 6312 S++ 0.48 -5.29 He 11.00 (0.02) 10.99 (0.01) 11.04 4686 He(++) 0.88 -2.19 6584 N+ 1.16 -5.72 N 6.28 (0.04) 6.94 (0.02) 6.96 7.63 4740 Ar'3 0.3: -6.16 6730 S+ 1.25 -6.27 O 7.97 (0.03) 8.46 (0.04) 8.57 8.79 Ne 7.40 (0.10) 7.83 (0.03) 7.99 7.86 S 6.5: 7.2: 7.2: 7.47 * Column (1) gives the wavelength in Angstroms; (2) the ion Ar 6.0: 6.1: 6.5: 5.95 whose abundance is computed from the line, for recombination helium lines this refers to next higher ionization stage; (3) the The figures in parentheses indicate the formal probable errors, logarithm of the intensity, on the scale I(H3) = 100, where (:) AlogN. For the Large Cloud, the values from model I are to be indicates an uncertain value; and (4) the logarithm of the ionic preferred. concentration calculated for model I. The derived logarithmic elemental abundances are for model I (and model II in paren- theses): He, -0.92 (-0.92); N, -5.12 (-5.05); 0, -3.68 tional inaccuracies. Our "straight mean" O/H ratio of 2.9 X (-3.3), Ne, -4.19 (-3.77)), S, -4.66 (4.63), Ar, -5.5 10-4 may err on the low side. Likewise, our N/H ratio may (-5.7). be too low if we systematically underestimate the n(N++) concentration. Since the ionization potential of Ne+, 41 eV, 4363/5007 intensity ratio as corrected for interstellar extinc- exceeds that of O+, 35 eV, the assumption one can estimate tion, and x from 6717/6731 ratio when possible. A weighted the concentration, n(Ne+) from a n(Ne++) value obtained mean abundance was found by assigning weights of 2 or 3 from [NeIII], X3868, and the n(O++)/n(O+) ratio may fail to better observed objects such as Henize 158, 159, 160, 44, except for the highest excitation objects such as H44C, which and especially 30 Doradus, and weights of l or even 1/2 for gives a Ne/H abundance ratio, 6.2 X 10-s, close to our the less well-observed objects such as Henize 8, 11, 51, 57, adopted value. The S/H abundance ratio is obtained from 30 59, 105, 119, 120, and 144. For model II, in which ad hloc Doradus, H44, and H160 where lines of both [SII] and [SIII] values were assigned to 1, we selected the eight best observed are easily measurable. Following the method of Peimbert objects. Fainter and less extensively observed nebulosities and Costero (11), we get n(S)/n(H) = 1.7 X 10-s. The un- gave mostly only N/H, O/H, and Ne/H ratios. For most certainty for argon is very large. Absence of [ArIn] X4740 sets elements, model II gives larger abundances than does model I a limit on n(Ar43), but concentrations in lower ionization and may represent limiting values, since here it is assumed stages are uncertain since X7135 [ArIIJ] is usually the only that temperature variations are so large that electron tem- observed line. The permitted X4267 CII line appears in Henize peratures found from the 4363/5007 ratio may apply only to 44, 158, and 30 Doradus, but its strength suggests that it is the very hottest portions of a given emitting volume and not excited directly by line radiation in the manner suggested by to an "average" zone where most of the quanta of a nebular- Kaler (43), rather than by recombination followed by cascade. type forbidden line actually originate. The data for the Small Magellanic Cloud are less extensive. The weighted mean He/H ratio, 0.098, based on 11 nebulos- Photoelectric measurements of the 4363/5007 ratio gave t = ities but without correction for possible incomplete helium 1.2 and 1.37 for NGC 346 (Henize 66) and Henize 83, respec- ionization exceeds the value found for the better observed tively. These values are in general accord with those reported 30 Doradus. Alternately, by applying Peimbert's procedure, by Dufour (44). For the remaining seven nebulosities studied, we found N(He+)/N(H+) = 0.096 ±t 0.006, so that there we assumed t = 1.2 for model I. The electron densities are appears to be no real discordance with the value obtained generally lower than in nebulosities of the Large Cloud. by the Peimberts. Further measurements are needed. Column 2 of Table 4 gives the mean chemical composition Henceforth, we adopt the model I abundances, which are of the Small Cloud, as derived exclusively from photoelectric based on temperatures determined from forbidden line ratios. measurements. Data for Henize 66 and 83 were given higher A limiting value of any element's abundance may be estimated weights than those for other objects. Nitrogen and oxygen as follows: If, as surmised in our first reconnaisance, the were observed in eight or nine nebulosities; other elements intrinsic chemical composition of the interstellar material in were observed in fewer objects. The probable errors again the Magellanic Clouds is everywhere the same, the population refer to the internal consistency of the measurements, not to ratio of a given ion with respect to hydrogen will depend only expected errors in final results. We examined implications on local levels of excitation and ionization. We might expect of two additional models: (a) t = 1.0, x = 0.01, and (b) model that in one or more regions of the Cloud, most of the atoms II with x = 0.01. Model (b) yielded considerably larger abun- of a particular element will exist in observable stages of ioniza- dances for N, 0, Ne, etc. Unfortunately, the helium abun- tion. This condition may occur often for helium and oxygen, dance is less well established. Available data for NGC 346 less frequently for sulfur and nitrogen, and perhaps only suggest n(He)/n(H) = 0.09 i 0.01, but further measurements rarely for argon and neon. Tyrpically,-the concentration [n(O+) need to be male. must very nearly equal the total oxygen abun- + n(O++)] SUMMARY AND DISCUSSION dance. From an examination of the individual n(ion)/n(H+) ratios for the best observed nebulosities, one might establish In accordance with earlier studies (45), we conclude that the an upper limit. There is the grave difficulty that abnormally helium/hydrogen ratio is probably similar in both Magellanic large values may represent only a bad choice of t or observa- Clouds. Certainly for the best observed object in the Large Downloaded by guest on September 23, 2021 Proc. Nat. Acad. Sci. USA 71 (1974) Composition of Nebulosities in Magellanic Clouds 4499

Cloud, 30 Doradus, our measured He/H ratio is smaller than 4. Wares, G. W., Ross, J. E. & Aller, L. H. (1968) Astrophys. in our own galaxy, in accord with the Peimberts' (33) excellent Space Sci. 2, 344-351. measurements. Results for other, less well-observed objects 5. Wolf, B. (1972) Astron. Astrophys. 20, 275-285. 6. Fry, M. & Aller, L. H. (1974) Ap. J. Suppl., in press. showed a larger scatter. We also agree with Dufour (44) that 7. Peimbert, M. & Spinrad, H. (1970) Astron. Astrophys. 7, nitrogen and oxygen are appreciably less abundant in the 311-317. Small Cloud than in the Large Cloud. Table 4 compares the 8. Peimbert, M. & Spinrad, H. (1970) Ap. J. 159, 809-815. adopted abundances for the two Magellanic Clouds with the 9. Peimbert, M. (1967) Ap. J. 150, 825-834. 10. Seaton, M. J. (1968) Mon. Not. Roy. Astron. Soc. 139, 129. results by Peimbert and Costero (11) (He, N, 0, Ne) and one 11. Peimbert, M. & Costero, R. (1969) Bol. Obs. Tonanzintla y of us (41) (S, Ar) for the Orion nebula, which can be taken as Tacubaya 3, Bull. no. 31. representative of the interstellar medium of our own galaxy. 12. Seraph, 1t. E. & Seaton, M. J. (1970) Mon. Not. Roy. Several trends are suggested. Excluding argon, for which the Astron Soc. 148, 367-381. 13. Krueger, T., Aller, L. H. & Czyzak, S. J. (1970) Ap. J. 160, Magellanic Cloud data involved large extrapolations, N, 0, 921-927. Ne, and S are less abundant than in the Orion nebula. The 14. Czyzak, S., Krueger, T. & Aller, L. H. (1970) Proc. Nat. depletion factor is small for the Large Cloud, probably less Acad. Sci. USA 66, 282-288. than a factor of two and therefore not striking for 0, Ne, and 15. Aller, L. H., Czyzak, S. J., Walker, M. P. & Krueger, T. K. S. Nitrogen would appear to be deficient a factor closer to (1970) Proc. Nat. Acad. Sci. USA 66, 1-5. by 16. Clarke, W. (1965) Dissertation, University of California, five, but again one must emphasize the uncertainty arising Los Angeles. from the extrapolation for this element. 17. Brocklehurst, M. (1971) Mon. Not. Roy. Astron. Soc. 153, In the Small Magellanic Cloud, the depletion of nitrogen, 471-490. 18. Henize, K. G. (1956) Astrophys. J. Suppl. 2, 315-344. oxygen and sulfur with respect to the Orion nebula is more 19. Doherty, L., Henize, K. G. & Aller, L. H. (1956) Ap. J. nearly an order of magnitude, while neon would appear to Suppl. 2, 345-363. be depleted by about half an order of magnitude. 20. Dickel, H. R., Aller, L. H. & Faulkner, D. J. (1964) Int. As long as the fluxes of ultraviolet radiation impinging Astron. Un. Symp. Nr. 20 on Galaxy and the Magellanic Clouds, eds. Kerr, F. J. & Rodgers, A. W. (Reidel, Dodrecht, upon the nebular gases are comparable in energy distribution, Holland), pp.. 294-310. the equilibrium electron temperature will depend on the 21. Aller, L. H. (1956 in Gaseous Nebulae (Chapman and Hall, efficacy of the energy dissipation by the excitation of for- London), p. 66. bidden lines (46). Thus, it is at least qualitatively under- 22. Dickel, H, R. (1965) Astrophys. J. 141, 1306-1317. 23. Mills, B. Y. & Aller, L. H. (1971) Aus. J. Phys. 24, 609-15. standable that the nebular plasma in the Small Cloud should 24. Mathewson, D. S. & Healey, J. R. (1964) Int. Astron. Un. be hotter than in the Large Cloud (44). Symp. Nr. 20, eds. Kerr, F. 1. & Rodgers, A. W. (Aus- The Large Megellanic Cloud would appear to represent a tralian Academy of Science, Canberra, Australia), pp. 245- system roughly comparable in chemical composition with our 255. 25. Clarke, J. N. (1971) Proc. Astron. Soc. Aus. 2, 44-45. own, although several investigators have suggested that the 26. McGee, R. X., Brooks, J. W. & Batchelor, R. A. (1972) stars may show a slightly smaller metal/hydrogen ratio than Aus. J. Phys. 25, 581-597, 613-617. our own galaxy, a conclusion consistent with our findings. 27. Broton, N. W. (1972) Aus. J. Phys. 25, 599-612. Likewise, Przybylski's (1) conclusion that the Small Cloud 28. McGee, R. X. & Newton, L. M. (1972) Aus. J. Phys. 25, in 619-635. may be deficient elements heavier than helium by about an 29. Faulkner, D. J. (1967) Mon. Not. Roy. Astron. Soc. 135, order of magnitude appears to be in harmony with results 401-412. obtained from the analysis of emission nebulosities. 30. Shapley, H. & Pareskevopoulos, J. S. (1937) Astrophys. J. 86, 340-341. 31. Feast, M. W. (1961) Mon. Not. Roy. Astron. Soc. 122, 1-16. We thank Director Blanco and the staff of Cerro Tololo Inter- 32. Johnson, H. M. (1959) Publ. Astron. Soc. Pac. 71, 425-434. american Observatory for helping us secure photographic and 33. Peimbert, M. & Torres-Peimbert, S. (1974) Astrophys. J., photoelectric measurements of the spectra of emission line in press. nebulosities in 1970, 1971, and 1973. This program was sup- 34. Mathis, J. S. (1965) Publ. Astron. Soc. Pac. 77, 189-201. ported in part by NSF Grant GP 31506X2 to UCLA and NSF 35. Faulkner, D. J. & Aller, L. H. (1965) Mon. Not. Roy. Grant GP 14601 to Ohio State University. The cooperation of the Astron. Soc. 130, 393-309. UCLA campus computing network is gratefully acknowledged. 36. Wares, G. W. & Aller, L. H. (1969) Publ. Astron. Soc. Pac. We thank the Peimberts for sending us their results in advance of 80, 568-577. publication. Dr. Jurgen Stock kindly supplied us unpublished 37. LeMarne, A. E. (1968) Mon. Not. Roy. Astron. Soc. 139, data on atmospheric extinction at Cerro Tololo that are in good 461-469. agreement with subsequently available determinations. We are 38. Nussbaumer, H. (1971) Ap. J. 166, 411-422. indebted to Harland Epps, who loaned us direct photographs of 39. Czyzak, S. J., Krueger, T. K., Martins, P., Seraph, H. E., the Magellanic Clouds secured with Curtis Schmidt. From these Seaton, M. J. & Shemming, J. (1968) Int. Astron. Un. we made finding charts to enable us to locate and return to well- Symp. 34 on Planetary Nebulae, eds. Osterbrock, D. E. & defined positions for spectroscopic and scanner observations. O'Dell, C. R. (Reidel, Dodrecht, Holland), pp. 138-142. 40. Seaton, M. J. (1974) Quart. J. Roy. Astron. Soc., in press. 1. Przybylski, A. (1968) Mon. Not. Roy. Astron. Soc. 139 41. Aller, L. H. (1972) Ann. N.Y. Acad. Sci. 194, 45-55. 313-339. 42. Bok, B. J. (1969) J. Roy. Astron. Soc. Can. 63, 105-24. 43. Kaler, J. B. (1972) Ap. J. 173, 601-609. 2. Przybylski, A. (1971) Mon. Not. Roy. Astron. Soc. 152, 44. Dufour, R. J. (1973) Bull. Amer. Astron. Sot. 5, 324. 197-208. 45. Aller, L. H. & Faulkner, D. J. (1962) Publ. Astron. Soc. 3. Przybylski, A. (1972) Mon. Not. Roy. Astron. Soc. 159, Pac. 74, 219-222. 155-163. 46. Menzel, D. H. & Aller, L. H. (1941) Astrophys. J. 94, 30-36. Downloaded by guest on September 23, 2021