1: \documentstyle[aaspp4]{article}
2:
3: %\received{ }
4: %\accepted{ }
5: %\journalid{ }{ }
6: %\articleid{ }{ }
7:
8: %\slugcomment{ }
9:
10:
11: \lefthead{Schneider \& Schombert}
12: \righthead{Nearby LSB Galaxies}
13:
14: \begin{document}
15:
16: \title{Nearby Gas-Rich Low Surface Brightness Galaxies}
17:
18: \author{Stephen E. Schneider}
19: \affil{Astronomy Program, University of Massachusetts,
20: Amherst, MA 01003}
21:
22: \author{James M. Schombert}
23: \affil{Department of Physics, University of Oregon}
24:
25: \begin{abstract}
26:
27: We examine the Fisher--Tully $cz<1000$ km s$^{-1}$ galaxy sample to determine
28: whether it is a complete and representative sample of {\it all} galaxy types,
29: including low surface brightness populations, as has been recently claimed.
30: We find that the sample is progressively more incomplete for galaxies with (1)
31: smaller physical diameters at a fixed isophote and (2) lower HI masses. This
32: is likely to lead to a significant undercounting of nearby gas-rich low
33: surface brightness galaxies. However, through comparisons to other samples we
34: can understand how the nearby galaxy counts need to be corrected, and we see
35: some indications of environmental effects that probably result from the local
36: high density of galaxies.
37:
38: \end{abstract}
39:
40: \keywords{galaxies: luminosity function}
41:
42: \section{Introduction}
43:
44: Two of the key questions in 20th century extragalactic studies have concerned
45: the density and the composition of the galaxy population in the Universe. Our
46: knowledge of galaxy types and their abundance depends critically on the issue
47: of completeness of our galaxy catalogs. Since our current catalogs are
48: constructed by various observational means, they are, by definition, limited by
49: natural and technological selection effects.
50:
51: Recently, Briggs (1997a) has argued that the Fisher \& Tully (1981, F--T)
52: catalog of nearby galaxies is complete to its redshift and sensitivity limits,
53: even for low surface brightness (LSB) galaxies. F--T examined HI emission from
54: a sample of 1787 angularly large galaxies accessible from the Green Bank radio
55: telescopes. They believed the sample to be very complete for late-type galaxies
56: within a redshift of $cz=1000$ km s$^{-1}$, with angular diameters larger than
57: 3$'$, located in regions at $|b|>30^\circ$ and $\delta>-33^\circ$. Briggs
58: additionally found a sensitivity limit depending on the HI mass and defined the
59: F--T ``completeness zone'' as extending out to:
60: \begin{equation}
61: z_{CZ} = \left\{ \begin{array}{ll} 1000\mbox{km s}^{-1}/c &
62: \mbox{if $M_{HI}>10^{8.45} h_{75}^{-2} M_\odot$} \\
63: (M_{HI}h_{75}^{2}/10^{8.45}M_\odot)^{5/12} 1000\mbox{km s}^{-1}/c &
64: \mbox{otherwise.}
65: \end{array} \right.
66: \end{equation}
67: The form of the limit for sources with masses smaller than
68: $10^{8.45}M_\odot$\footnote{We adopt $H_0=75$ km s$^{-1}$ Mpc$^{-1}$ for masses
69: and distances quoted hereafter, and we use the heliocentric velocity corrected
70: by 300 km s$^{-1} \cos b \sin l$ as in F--T and Briggs.} was based on a
71: semi-analytic, semi-empirical fit to the HI sensitivity.
72:
73: Briggs pointed out that surveys of low surface brightness galaxies (Schneider
74: et al.~1990; Schombert et al.~1992; Matthews \& Gallagher 1996; Impey et
75: al.~1996) have identified sources that are primarily at larger distances, but
76: they have added very few within the F--T ``completeness zone.'' He concluded
77: that LSB galaxies ``must already be fairly represented by nearby, previously
78: cataloged examples.''
79:
80: This is an interesting idea, but the conclusion does not necessarily follow
81: from the analysis for several reasons: (1) For sources with low HI masses
82: $z_{CZ}$ is so small that redshift distances are very uncertain and Galactic
83: HI emission creates strong confusion. A further objection to Briggs' analysis
84: is that (2) the F--T sources are themselves incomplete within Briggs'
85: ``completeness zone'' because of angular size selection effects. Lastly, (3)
86: the $cz<1000$ km s$^{-1}$ region around the Local Group has about twice the
87: average galactic density and therefore is not a very representive sample of the
88: Universe as a whole.
89:
90: In a companion paper, Briggs (1997b) used the F--T and LSB samples to derive
91: an HI mass function. He commented on the need for corrections for
92: incompleteness and noted the usefulness of a ${\cal V}/{\cal V}_{max}$ test
93: for establishing whether the LSB samples were complete, but he did not discuss
94: the problems the F--T sample has in this regard. In this paper we explore the
95: limitations of the F--T sample and discuss how it and more recent surveys may
96: be properly used to understand the composition of galaxy populations. We show
97: that the local samples of galaxies display morphological segregation
98: characteristics associated with high density environments. Finally, we find
99: that in all HI mass ranges the dominant class of galaxies are those with the
100: smallest angular diameters at the isophotal limits of the original Palomar Sky
101: Survey (PSS-I). These galaxies may be physically small or appear small because
102: they are LSB systems; in either case, they are greatly under-represented in
103: the F--T sample and most other optical surveys.
104:
105:
106: \section{Completeness Tests of the F--T Sample}
107:
108: The incompleteness of the F--T sample can be demonstrated using a ${\cal V} /
109: {\cal V}_{max}$ test. This test compares the distance $d$ of a detected source
110: to the maximum distance $d_{max}$ at which it should be detectable. If the
111: maximum distance is correctly estimated, a source is equally likely to fall
112: anywhere within the volume delimited by $d_{max}$. On average, then, sources
113: will be found halfway into the maximum volume, and ${\cal V} / {\cal V}_{max}
114: \equiv (d/d_{max})^3$ will average 0.5 (Schmidt 1968). For a sample of $N$
115: sources, the probability distribution of the mean value of ${\cal V}/{\cal
116: V}_{max}$ has a nearly normal distribution with standard deviation
117: $1/\sqrt{12N}$.
118:
119: We assume the distance is proportional to the redshift $z$, so that ${\cal
120: V}/{\cal V}_{max}=(z/z_{max})^3$. We exclude galaxies within $6^\circ$ of the
121: center of the Virgo cluster or $3^\circ$ of the Fornax cluster from the
122: $cz<1000$ km s$^{-1}$ sample. This eliminates the worst distance estimates,
123: although peculiar velocities clearly must affect the rest of the sample as
124: well. We have tested how adjustments for peculiar velocity might alter our
125: results using the $POTENT$ model of Dekel, Bertschinger, and Faber (1990), and
126: find no substantive changes from the results presented below, although the
127: samples generally have somewhat lower values of ${\cal V}/{\cal V}_{max}$. We
128: use the redshift corrected for Local Group motion here to maintain consistency
129: with F--T and Briggs.
130:
131: With $z_{max}=z_{CZ}$, the mean value of ${\cal V}/{\cal V}_{max}$ is
132: $0.406\pm0.016$. Values below 0.5 imply that galaxies were detected
133: preferentially in the nearer portion of the survey volume, suggesting that the
134: sample is not fully sensitive to sources out to $z_{CZ}$.
135:
136: A low value of ${\cal V}/{\cal V}_{max}$ can alternatively be caused by an
137: actual clustering of galaxies nearby us. However, this tends to be
138: counter-balanced by the effects of morphological segregation and gas
139: depletion, which would favor HI detections in lower-density environments. In
140: any case, the F--T sample is mostly confined to within the local supercluster,
141: and it is not clear that there is an overall radial gradient within the
142: sampled region. Moreover, these effects do not explain the dependence of
143: ${\cal V}/{\cal V}_{max}$ on HI mass.
144:
145: \input tbl1.tex
146:
147:
148: In Table 1 we list the results for various galaxy samples and divide each
149: sample into three mass ranges.\footnote{While the upper and lower mass ranges
150: are unbounded on one side, the range of detected masses is approximately one
151: decade in both cases.} Sample 1 shows that the F--T sample in the
152: ``completeness zone'' exhibits worse and worse completeness for lower mass
153: galaxies.
154:
155: Using Briggs' functional form (eqn.~1) for the completeness limit, we could
156: increase the minimum mass for full-volume sensitivity from $10^{8.45}$ to
157: $10^{9.15} M_\odot$ to make ${\cal V}/{\cal V}_{max}$ close to 0.5 in all mass
158: ranges (sample 2 in Table 1). This revised limit would set the completeness
159: zone limit to $cz<330$ km s$^{-1}$ for $M_{HI}<10^8$ so that distance
160: uncertainty and confusion with Galactic HI would present significant problem
161: for an even larger portion of the sample. In addition, within such small
162: redshifts ${\cal V}/{\cal V}_{max}$ is probably biased upward, since the
163: lowest redshift sources may be lost in Galactic emission and sources detected
164: at redshifts below $cz=100$ km s$^{-1}$ were set to that value for the purpose
165: of estimating their distances. These limitations make the F--T sample highly
166: problematic for trying to understand properties of galaxies with HI masses
167: $<10^8 M_\odot$.
168:
169:
170: \section{Angular Size Limitations of Optical Samples}
171:
172: Another approach to understanding the completeness of the F--T sample is to
173: examine the source selection criteria. F--T used a minimum angular size as
174: their primary selection criterion---examining spirals with diameters $a>3'$ and
175: Sd--Im galaxies with $a>2'$ as determined in the UGC (Nilson 1973). Because
176: other sources of angular diameters were also used for parts of the sample, some
177: smaller galaxies were also observed. To place all of the angular sizes on a
178: common system, we use the formulas from F--T to convert to the UGC scale.
179:
180: Based on the expected increase of counts with angular diameter as $N\propto
181: a^{-3}$, the full sample of observed galaxies (whether or not they were
182: detected in HI) begins to be incomplete for angular sizes $a<4'$
183: (Fig.~\ref{aftu}). At $a=2'$ there are $\sim15\times$ too few galaxies relative
184: to the larger sources. This incompleteness at small angular sizes is partly
185: intentional, since F--T excluded small angular diameter galaxies that they
186: expected would be distant. Unfortunately, this also introduces a degree of
187: subjectiveness to inclusion in the sample.
188:
189: \begin{figure}[tbh]
190: \plotone{fig1.eps}
191: \caption{
192: Histogram of F--T galaxy angular diameters. The dotted line shows the expected
193: slope for a complete sample distributed uniformly throughout space.
194: }
195: \label{aftu}
196: \end{figure}
197:
198: Even among the F--T galaxies in Briggs' ``completeness zone,'' many of the
199: galaxies have angular sizes so small that they would not have remained in the
200: sample if they were at their maximum distance within the zone. The angular size
201: a source would have at $z_{CZ}$ is $a_{CZ}\equiv a\times z/z_{CZ}$. Of the 41
202: sources with $M_{HI}<10^8M_\odot$, none has $a_{CZ}>4'$; only one is $>3'$, the
203: stated completeness limit of F--T; and only 8 have $a_{CZ}>2'$. Even among the
204: 171 intermediate mass sources, with $10^8<M_{HI}<10^9M_\odot$, only 7\%, 25\%,
205: and 60\% galaxies have $a_{CZ}>4'$, $3'$, and $2'$ respectively. Only the high
206: mass sources are large enough that high fractions pass the $a_{CZ}$
207: requirement---74\%, 90\%, and 98\% for $a_{CZ}>4'$, $3'$ and $2'$. Clearly this
208: will tend to push ${\cal V}/{\cal V}_{max}$ to lower values since some galaxies
209: are included only in the near portion of the search volume.
210:
211: To make a more uniform selection we can restrict the F--T sample to $a_{CZ}>3'$
212: (sample 3), in the intermediate and high mass ranges ${\cal V}/{\cal V}_{max}$
213: is below 0.5 only marginally (1.3 and 1.0 $\sigma$), but the low mass range
214: cannot be tested since it has only one galaxy. Restricted to $a_{CZ}>2'$
215: (sample 4) the F--T sample does not pass the ${\cal V}/{\cal V}_{max}$ test in
216: either the low or intermediate mass ranges.
217:
218: One way of addressing the omission of small galaxies is to look to samples of
219: small galaxies. In particular the ``dwarf and LSB'' (D+LSB) sample of galaxies
220: from Schneider et al. (1990, 1992) contains HI measurements for late-type,
221: dwarf, and irregular galaxies down to a $1'$ diameter. Since this sample is
222: drawn from the UGC, it covers only the northern sky ($\delta>-2.5^\circ$), but
223: when supplemented with HI measurements from the literature (Huchtmeier \&
224: Richter 1989) the HI detections are more than 85\% complete.
225:
226: The ${\cal V}/{\cal V}_{max}$ test results for the D+LSB sample within Briggs'
227: ``completeness zone'' are given in Table 1 (sample 5). The low-mass ranges
228: still do not pass the test, but they fare considerably better than the F--T
229: sample. By restricting the galaxies to $a_{CZ}>1'$, which eliminates galaxies
230: that only meet the UGC size criterion because they are very nearby (sample 6),
231: the test is passed to within $2\sigma$ in all mass ranges. This also shows
232: that large scale structure is not causing low ${\cal V}/{\cal V}_{max}$ test
233: results for the F--T sample.
234:
235: We can combine the F--T and D+LSB samples in the hope of forming a complete
236: sample of all types of galaxies as Briggs (1997b) did. In the northern sky we
237: find 248 F--T galaxies and 47 additional D+LSB galaxies that satisfy the
238: ``completeness zone'' criteria. This expanded sample fares only marginally
239: better in the ${\cal V}/{\cal V}_{max}$ tests, yielding $0.407\pm0.016$ for the
240: full sample, and in the separate mass ranges (Table 1, sample 7). Restricting
241: $a_{CZ}$ does not generate samples that pass the ${\cal V}/{\cal V}_{max}$ test
242: either.\footnote{Restricting the samples to high Galactic latitudes
243: ($|b|>30^\circ$) made no appreciable difference to the results presented in
244: Table 1.}
245:
246: We conclude that the F--T sources with high HI masses represent a relatively
247: complete sample, but the sources with low HI masses are strongly biased to low
248: redshifts. The problem with low mass galaxies may be caused in part by the
249: F--T angular size criterion. However, even when the F--T sample is (1)
250: restricted to minimum physical diameters to make the galaxies relatively
251: uniform within the ``completeness zone,'' or (2) supplemented by galaxies from
252: other surveys, the galaxies with HI masses $<10^9 M_\odot$ still fail the
253: ${\cal V}/{\cal V}_{max}$ test. This demonstrates that the F--T sample does
254: not provide a good basis for forming a representative cross section of galaxy
255: types.
256:
257: \section{High Mass LSB Galaxies}
258:
259: While galaxies with high HI masses ($M_{HI}>10^9 M_\odot$) in the F--T sample
260: pass the ${\cal V}/{\cal V}_{max}$ test, this is really only an internal check
261: on the self-consistency of the database. To examine the broader question of
262: how representative the F--T sample is, Briggs (1997a) asked whether surveys of
263: LSB galaxies had found galaxies within the ``completeness zone.'' However,
264: since these other surveys were also based on visual examination of
265: photographic plates, they do not provide a genuinely independent check of the
266: F--T sample. Moreover, since the local density of galaxies is higher than
267: average, classes of galaxies that avoid high density may not be present.
268:
269: The question we consider here is whether there are massive HI sources in
270: deeper surveys that would have been excluded from the F--T sample because
271: their isophotal diameters at the PSS-I surface brightness would be less than
272: 3--4$'$ at the 1000 km s$^{-1}$ redshift limit for high mass galaxies in the
273: ``completeness zone.'' This is difficult to quantify precisely since diameters
274: estimated from the PSS-I are somewhat variable in their effective depth, but
275: we will adopt the mean isophotal level found by Cornell et al.~(1987) of
276: $\mu_{PSS-I}\equiv 25.36$ mag arcsec$^{-2}$ at $B$ for UGC diameters.
277:
278: The F--T subset of high-mass galaxies are almost all physically large at
279: $\mu_{PSS-I}$. Figure 2 shows the distribution of $a_{CZ}$ for this (solid-line
280: histogram) and other samples of galaxies. Since all of the high mass galaxies
281: are detectable to the 1000 km s$^{-1}$ redshift limit, $a_{CZ}=1'$ corresponds
282: to 3.88 kpc. Thus 90\% of the F--T high mass galaxies have sizes larger than
283: 11.6 kpc. Note that we restrict the following analyses to galaxies at high
284: Galactic latitudes where interstellar extinction should not much affect the
285: galaxies' optical sizes or number counts. \footnote{Briggs (1997ab) specifies
286: that his samples are restricted to high latitudes, but the numbers of galaxies
287: he quotes in various subsamples indicate he was using the full sky coverage
288: of F--T.}
289:
290: \begin{figure}[p]
291: \epsscale{0.7}
292: \plotone{fig2.eps}
293: \caption{
294: Number density of galaxies as a function of their angular diameter as
295: determined at the ``completeness zone'' distance limit. The size distribution
296: of the F--T sample is shown by a solid-line histogram; the UGC Dwarf+LSB
297: sample is shown by a dashed line, and restricted to the ``completeness zone''
298: by a dot-dash line; the PSS-II LSB sample is shown by a dotted line. The size
299: distribution of the HI-selected ``Slice'' sample is shown by a solid gray
300: histogram. The densities of the F--T sample and the D+LSB sample in the
301: low-mass range are divided by two to account for the local overdensity as
302: explained in the text.
303: }
304: \label{massacz}
305: \end{figure}
306:
307:
308:
309: We give the estimated number density of each size of galaxy in the figure
310: based on the areal coverage of the sample ($\sim$5.1 sr for the high-latitude
311: portion of the F--T sample). We also need to account for the local overdensity
312: of galaxies when making comparisons to other samples. Briggs (1997b) estimates
313: the region inside $cz<1000$ km s$^{-1}$ has a density about a factor of 2
314: above average. This matches our results for the D+LSB sample, which has a
315: density 2.1 times higher in the nearby portion. Densities in Fig.~2 that are
316: based on the F--T sample and other samples restricted to the local region are
317: divided by 2.
318:
319: The ``HI-Slice'' sample of Spitzak \& Schneider (1998) was found by
320: systematically observing 55 sq deg of the sky at 21 cm from Arecibo, and is
321: unbiased by optical sizes. This survey contains 62 sources with $M_{HI}>10^9
322: M_\odot$. We have determined the angular sizes at $\mu_{PSS-I}$ from the
323: original $B$-band photometric profile fits of Spitzak \& Schneider. Compared to
324: the F--T sample, these galaxies have lower percentages of optically large
325: galaxies, and higher percentages of small galaxies, although there appear to be
326: very few galaxies smaller than 7.8 kpc in either sample.
327:
328: Using the sensitivity limit estimates from Schneider, Spitzak, \& Rosenberg
329: (1998), we can determine the volume within which each of the HI-Slice galaxies
330: was detectable in order to estimate its number density in space. The results
331: are shown in Fig.~2 by the solid-gray histogram. We estimate that 75\% of the
332: HI among these massive galaxies is associated with galaxies larger than
333: $a_{CZ}>4'$. Assuming the F--T sample is complete for these largest galaxies,
334: the smaller fraction of galaxies it finds at smaller sizes implies it is
335: missing about 23\% of high-mass galaxies and 12\% of the total HI due to the
336: angular size limitations.
337:
338: Several differences between the samples may reflect environmental influences.
339: The distribution of F--T galaxy sizes in Fig.~2 suggests that the population
340: has been shifted to systematically larger sizes than the HI-Slice galaxies. In
341: addition, the HI-selected sample has $1.8\times$ less $B$-band luminosity
342: relative to the HI mass on average, based on optical data from Spitzak \&
343: Schneider and the RC3 (de Vaucouleurs et al. 1991), and 40\% of the HI-selected
344: galaxies have $M_{HI}/L_{B}>1$ (in solar units) compared to 11\% of the F--T
345: galaxies. Since the F--T sample is located in a region of high galaxy
346: density, star formation induced by galaxy interactions may explain these
347: differences.
348:
349: By contrast, {\it optically}-selected samples, even of LSB galaxies, rarely
350: identify high HI mass objects that would not have been identified in Briggs'
351: ``completeness zone.'' If we impose no HI-mass/redshift restrictions, the
352: D+LSB sample (\S 3) contains 586 high-HI-mass sources with $|b|>30^\circ$, and
353: it has even higher fractions of large-$a_{CZ}$ galaxies than the F--T sample.
354: This suggests that almost all UGC galaxies with large HI masses have large
355: physical dimensions at $\mu_{PSS-I}$.
356:
357: However, the raw distribution of $a_{CZ}$ in an angular-diameter limited
358: sample is not a good indicator of the population size distribution since
359: galaxies with smaller physical diameters at $\mu_{PSS-I}$ remain larger than
360: $1'$ (the UGC limit) out to smaller distances. Dividing the counts by the
361: volume in which each galaxy remains larger than $1'$, the distribution has a
362: similar shape to the HI-selected sample (dashed-line histogram in Figure 2).
363: These galaxies also appear to have distinctly different properties from
364: galaxies in the F--T sample---their $B$-band luminosities relative to their HI
365: masses are similar to the HI-selected sample of sources.
366:
367: The distribution of sizes of the subset of D+LSB galaxies within $cz<1000$ km
368: s$^{-1}$ is shown by a dot-dash line. The counts have been divided by 2 (like
369: the F--T sample) and demonstrate that the density correction used earlier is
370: reasonable. The nearby D+LSB galaxies show a slight shift toward larger sizes
371: than the full sample. This indicates that the distribution of sizes in a
372: sample of late-type galaxies alone is little-affected by the local density
373: enhancement. By contrast, the F--T and HI-selected samples contain a wide
374: range of morphological types. Since earlier-type galaxies tend to be larger at
375: the same isophotal level, the shift in the size distribution to larger
376: $a_{CZ}$ within the F--T sample may reflect the effects of morphological
377: segregation, with a larger proportion of earlier types than an average sample.
378:
379: Within the northern portion of the ``completeness zone'' the D+LSB sample adds
380: only 1 high-mass galaxy to the F--T sample, so the size distribution of the
381: combined sample is basically unchanged. The lack of additional
382: small angular diameter galaxies supports the idea that the local population is
383: different from samples drawn from a wider variety of environments.
384:
385: The PSS-II survey of Eder \& Schombert (1999) is a deeper probe for small
386: diameter, very late-type galaxies. These objects were selected from 50 PSS-II
387: plates ($\sim$0.7 sr) to be larger than 20$''$ and to have dwarf-like
388: morphologies. We have determined the sizes on the PSS-I for each of these
389: objects; our size estimates are consistent for galaxies in common with the UGC.
390: The size distribution turns over below diameters of 0.4$'$ at $\mu_{PSS-I}$,
391: which we take to be the effective completeness limit.
392:
393: While aimed primarily at identifying dwarf galaxies, the PSS-II sample includes
394: 135 galaxies with $M_{HI}>10^9M_\odot$. Their size distribution is shown by a
395: dotted line in Fig.~2. Only 7 of these objects are smaller than $a_{CZ}<2'$,
396: but based on their angular size distance limits, such small sources comprise
397: about half of the population of high mass LSB objects with very late type
398: morphologies.
399:
400: Extremely high HI mass LSB systems like Malin 1 (Bothun et al.~1987) might
401: have been missed in the nearby volume of space because their disks are so
402: faint that even the central extrapolated surface brightness of the disk is
403: fainter than $\mu_{PSS-I}$. However, the bulge component of Malin 1 would
404: reach $\mu_{PSS-I}$ at about $3.75'$ at $cz=1000$ km s$^{-1}$. We can
405: speculate that a system like Malin 1 might have been identified as an E or S0
406: in the UGC, and therefore omitted from consideration for the F--T sample. It
407: is worth noting that a number of early-type galaxies have been found with
408: extended distributions of HI (Van Driel \& Van Woerden 1991; DuPrie \&
409: Schneider 1996), and perhaps these are the more appropriate comparison to
410: Malin 1. For other giant disk systems that have been compared to Malin 1, like
411: F568-6 (Bothun et al.~1990) and 1226+0105 (Sprayberry et al.~1993), the disk
412: surface brightness and size would make these objects easily exceed
413: $a_{CZ}>3'$, and should thus have been included in the F--T sample if they
414: were in the ``completeness zone.''
415:
416: In summary, high-HI-mass galaxies are relatively well-sampled by F--T, but
417: they do miss an interesting fraction of galaxies that have small sizes at the
418: PSS-I isophotal limit. Based on intrinsic differences in the optical-to-21 cm
419: emission from the nearby F--T sample versus more-distant samples, we suggest
420: that the high density of galaxies in the local environment may cause
421: differences in the local population.
422:
423: \section{Low and Intermediate Mass Galaxies}
424:
425: Galaxies with less than $10^9 M_\odot$ of HI are highly incomplete in the F--T
426: sample. Among the galaxies identified by F--T in this mass range, about half
427: have $a_{CZ}<2'$, and only $7\%$ are larger than $a_{CZ}>4'$, so F--T's
428: adopted angular size constraints give rise to a fundamental limitation to the
429: sample's completeness. We illustrate here the degree of the incompleteness and
430: attempt to extrapolate to the population of missing objects.
431:
432: The distribution of sizes in the F--T sample among lower mass galaxies is shown
433: in the bottom two panels of Fig.~2, divided into intermediate
434: ($10^8-10^9M_\odot$) and low ($<10^7M_\odot$) HI masses. Since $z_{CZ}$
435: declines for galaxies with HI masses $<10^{8.45}M_\odot$ (eqn.~1), the density
436: is estimated from the corresponding volume. At $10^8 M_\odot$, the distance
437: limit and $a_{CZ}$ are 65\% of their value for high-mass galaxies, and at $10^7
438: M_\odot$ they are 25\% as big.
439:
440: The HI-Slice sample of Spitzak \& Schneider (1998) contains 3 low and 10
441: intermediate mass galaxies. Despite the small-number statistics, this sample
442: clearly demonstrates that very small sizes are the norm among low-HI-mass
443: galaxies as shown in Fig.~2. {\it All} of the sources are smaller than
444: $a_{CZ}=2.2'$, and 5 of 6 sources with $M_{HI}<2.5\times10^8$ are smaller than
445: $a_{CZ}<1'$. One relatively high mass source ($M_{HI}=7\times10^8M_\odot$) has
446: $a_{CZ}=0.13'$ and is nearly invisible on the PSS-I.
447:
448: The size-distribution of the HI-Slice galaxies is clearly different from the
449: F--T sample, exhibiting a strong peak toward the smallest diameters. The
450: lowest-mass source in the HI-Slice sample (\#75 in Spitzak \& Schneider) was
451: not detected optically because of a foreground star, but it is certainly
452: very small. Given its low potential detection volume, it would increase the
453: estimated density of the smallest-size low-mass galaxies by more than a factor
454: of six. Since its contribution is not included in the histogram, the density of
455: very small low-mass galaxies may be substantially larger than shown.
456:
457: The D+LSB sample contains 482 galaxies with $M_{HI} < 10^9 M_\odot$. 22 of
458: these galaxies have $a_{CZ}<1'$, ranging down to $0.44'$ and physical diameters
459: as small as 0.7 kpc at the UGC isophote. All 22 of these small galaxies fall
460: within the F--T ``completeness zone'' but such galaxies would be overlooked
461: even in the combined F--T and D+LSB samples if they were beyond the nearest
462: portion of the zone.
463:
464: Most of the D+LSB galaxies are larger than $2'$, but after adjusting for each
465: source's maximum detectable distance according to its angular size, we find the
466: density distribution shown by the dashed-line histogram in Fig.~2. In the
467: intermediate mass range this distribution is quite similar to the HI-Slice
468: sample for galaxies with small diameters.
469:
470: The portion of the D+LSB sample restricted to the ``completeness zone'' (24\%
471: of 394 galaxies) is again shown with a dot-dash histogram in Fig.~2. The size
472: distribution of these galaxies begins to resemble that of the F--T sample even
473: though small-$a_{CZ}$ galaxies would be easier to detect nearby. This suggests
474: again that the nearby volume of space is atypical.
475:
476: All three of the low-mass HI-Slice galaxies and 92\% of the 88 low-mass D+LSB
477: galaxies have redshifts below $cz=1000$ km s$^{-1}$, although most are outside
478: the ``completeness zone'' distance limit at these masses. We have divided the
479: densities of the low-mass D+LSB galaxies by a factor of 2 as we did for the
480: F--T sample since it mostly probes the same volume of space. The HI-Slice
481: sample density estimates are already adjusted for the local large-scale
482: structure in the direction of that survey (Schneider et al. 1998).
483:
484: We estimate the space density of small, LSB galaxies based on 23 low-mass and
485: 77 intermediate-mass galaxies that are larger than $0.4'$ in the PSS-II sample
486: of Eder \& Schombert (1999). Most of the low-mass galaxies are within $cz<1000$
487: km s$^{-1}$, so the densities should perhaps be divided by 2; however,
488: the sample is partially incomplete for the smallest sizes, so the densities may
489: be underestimated. The densities (dotted histogram in Fig.~2) assume that
490: sources were detectable out to the distance where their angular diameter would
491: reach $0.4'$.
492:
493: The lowest-mass galaxy in the Eder \& Schombert LSB sample (D634-3) was not
494: included in the density estimates. For this galaxy $V_0=181$ km s$^{-1}$, so
495: its distance and mass are quite uncertain. Taken at face value this source
496: would imply a density of 0.7 Mpc$^{-3}$ of very low-mass objects, comparable to
497: the large density implied by the lowest-mass source in the HI-Slice sample.
498:
499: In summary, optically-selected samples of galaxies only begin to indicate the
500: prevalance of small-diameter galaxies as measured at the limiting isophotal
501: depth of the PSS-I. Photographic surveys of galaxies with late-type
502: morphologies can recover the density of objects with intermediate HI masses if
503: the selection criteria are well-understood, but low mass galaxies present a
504: much bigger challenge. Detections of two very small, low mass galaxies in an HI
505: survey and an LSB survey imply that there may be a very large population of
506: sources with $M_{HI}<10^7M_\odot$, but the statistical uncertainties are too
507: great to draw firm conclusions on this point.
508:
509: \section{Discussion}
510:
511: Optically selected samples favor optically bright galaxies. This truism holds
512: even for diameter-limited galaxy surveys because LSB galaxies appear small at
513: the surface-brightness limit of the optical images (Disney 1976). The HI-Slice
514: survey (Spitzak \& Schneider 1998), which is unbiased by galaxy diameter,
515: indicates that the optically smallest galaxies are the most common. Current
516: searches on deep photographic plates for small angular diameter sources (Eder
517: \& Schombert 1999) are also uncovering indications of this population. Such
518: LSB and 21 cm surveys are successfully probing sources with HI masses down to
519: $\sim10^8M_\odot$, but for lower mass objects HI flux and angular size
520: limitations of existing surveys allow detections of these sources to only a
521: few Mpc.
522:
523: Because of the small distances at which low mass and LSB sources are
524: accessible, we need to consider the possible impact of the local environment
525: on them. Although the large scale distribution of LSB galaxies appears similar
526: to that of other galaxies (Mo, McGaugh and Bothun 1994), on scales of less
527: than 2 Mpc their numbers drop off sharply. The most likely explanations are
528: that either LSB disks are fragile and easily disturbed by other galaxies, or
529: tidal interactions induce star formation that converts LSB galaxies into
530: normal Hubble type objects. Regardless of the underlying cause why LSB
531: galaxies avoid high density regions, this fact produces an expectation that
532: the local region of space will be deficient in the number of LSB disk galaxies
533: due to the large number of high mass spirals and the proximity to the very
534: dense Virgo Cluster.
535:
536: Other influences of the local environment may also play a role in the
537: distribution of galaxy types. In high density regions, galaxies are often gas
538: deficient for their morphological type because of stripping or evaporation,
539: and morphological segregation favors earlier-type, less gas rich galaxies.
540: Both of these effects would tend to {\it lower} the percentage of gas-rich
541: systems nearby.
542:
543: Briggs (1997a) asked ``Where are the nearby gas-rich low surface brightness
544: galaxies?'' The answer depends on the mass range of objects being studied.
545: There appears to be a local deficit of high-mass LSB systems, which is
546: probably an environmental effect. The story for low-mass systems is less
547: settled, because of the difficulty in detecting them to any significant
548: distance, but it is clear there is a much larger population of
549: small-optical-diameter galaxies than optical surveys have previously revealed.
550: And finally, since these low-mass objects have not yet been detected beyond
551: the local high-density environment, it is possible that they are even more
552: profuse than they appear locally.
553:
554: \begin{acknowledgments}
555:
556: \end{acknowledgments}
557:
558: \begin{thebibliography}{}
559:
560: \bibitem{}
561: Bothun, G. D., Impey, C. D., Malin, D. F., Mould, J. R. 1987, AJ, 94, 23
562:
563: \bibitem{}
564: Bothun, G. D., Schombert, J. M., Impey, C. D., \& Schneider, S. E. 1990, ApJ, 360, 427
565:
566: \bibitem{}
567: Briggs, F. H. 1997a, ApJ, 484, L29
568:
569: \bibitem{}
570: Briggs, F. H. 1997b, ApJ, 484, 618
571:
572: \bibitem{}
573: Cornell, M. E., Aaronson, M., Bothun, G., \& Mould, J. 1987, ApJS, 64, 507
574:
575: \bibitem{}
576: Dekel, A., Bertschinger, E., \& Faber, S. M. 1990, ApJ, 364, 349
577:
578: \bibitem{}
579: de Vaucouleurs, G., de Vaucouleurs, A., Corwin, H. C., Buta, R. J., Paturel,
580: G., \& Fouque, P. 1991, Third Reference Catalog of Bright Galaxies (New York:
581: Springer)
582:
583: \bibitem{}
584: Disney, M. J. 1976, Nature, 263, 573
585:
586: \bibitem{}
587: DuPrie, K., \& Schneider, S. E. 1996, AJ, 112, 937
588:
589: \bibitem{}
590: Eder, J. A., \& Schombert, J. M. 1999, in preparation
591:
592: \bibitem{}
593: Fisher, J. R., \& Tully, R. B. 1981, ApJS, 47, 139
594:
595: \bibitem{}
596: Huchtmeier, W. K., \& Richter, O. G. 1989, {\it A General Catalog of HI
597: Observations of Galaxies} (Berlin: Springer-Verlag)
598:
599: \bibitem{}
600: Impey, C. D., Sprayberry, D., Irwin, M. J., \& Bothun, G. D. 1996, ApJS, 105, 209
601:
602: \bibitem{}
603: Matthews, L. D., \& Gallagher, J. S. 1996, AJ, 111, 1098
604:
605: \bibitem{}
606: Mo, H. J., McGaugh, S. S., \& Bothun, G. D. 1994, MNRAS, 267, 129
607:
608: \bibitem{}
609: Nilson, P. 1973, Uppsala Astron. Obs. Ann. 6 (UGC)
610:
611: \bibitem{}
612: Schmidt, M. 1968, ApJ, 151, 393
613:
614: \bibitem{}
615: Schneider, S. E., Spitzak, J. G., \& Rosenberg, J. L. 1998, ApJ, 507, L9
616:
617: \bibitem{}
618: Schneider, S. E., Thuan, T. X., Magri, C., \& Wadiak, J. E. 1990, ApJS, 72, 245
619:
620: \bibitem{}
621: Schneider, S. E., Thuan, T. X., Mangum, J. G., \& Miller, J. 1992, ApJS, 81, 5
622:
623: \bibitem{}
624: Schombert, J. M., Bothun, G. D., Schneider, S. E., McGaugh, S. S. 1992, AJ, 103, 1107
625:
626: \bibitem{}
627: Sprayberry, D., Impey, C. D., Irwin, M. J., McMahon, R. G., \& Bothun, G. D.
628: 1993, ApJ, 417, 114
629:
630: \bibitem{}
631: Spitzak, J. G., \& Schneider, S. E. 1998, ApJS, 119, 159
632:
633: \bibitem{}
634: Van Driel, W., \& Van Woerden, H. 1991, \aap, 243, 71
635:
636: \end{thebibliography}
637:
638: \end{document}
639:
640: \end
641:
642: