1: \documentclass[12pt,preprint]{aastex}
2:
3: \begin{document}
4:
5: \title{Rapid Formation of Super-Earths around M Dwarf Stars}
6:
7: \author{Alan P.~Boss}
8: \affil{Department of Terrestrial Magnetism, Carnegie Institution of
9: Washington, 5241 Broad Branch Road, NW, Washington, DC 20015-1305}
10: \authoremail{boss@dtm.ciw.edu}
11:
12: \begin{abstract}
13:
14: While the recent microlensing discoveries of super-Earths orbiting two
15: M dwarf stars have been taken as support for the core accretion mechanism
16: of giant planet formation, we show here that these planets could
17: also have been formed by the competing mechanism of disk instability,
18: coupled with photoevaporative loss of their gaseous envelopes by a
19: strong external source of UV radiation, i.e., an O star. M dwarfs that
20: form in regions of future high-mass star formation would then be expected
21: to have super-Earths orbiting at distances of several AU and beyond,
22: while those that form in regions of low-mass star formation would
23: be expected to have gas giants at those distances. Given that most
24: stars are born in the former rather than in the latter regions, M
25: dwarfs should have significantly more super-Earths than gas giants,
26: as seems to be indicated by the microlensing surveys.
27:
28: \end{abstract}
29:
30: \keywords{stars: planetary systems -- stars: low-mass, brown dwarfs}
31:
32: \section{Introduction}
33:
34: Microlensing surveys have discovered recently two ``super-Earths''
35: orbiting M dwarf stars, planets with masses of $\sim 5.5 M_\oplus$
36: (Beaulieu et al. 2006) and $\sim 13 M_\oplus$ (Gould et al. 2006).
37: As microlensing detections yield only the ratio of the lensing planet mass
38: to the lensing star mass, these planetary mass estimates rely heavily
39: on the unproven assumption that the lensing star is an M dwarf. The
40: super-Earths have been interpreted as being ``failed cores'' produced
41: by the first step of the core accretion mechanism for gas giant
42: planet formation. The super-Earths presumably failed to become
43: gas giant planets because the growth of solid cores by collisional
44: accumulation proceeds considerably slower at a fixed orbital radius
45: around an M dwarf star than around a G dwarf star (Laughlin, Bodenheimer,
46: \& Adams 2004), and the disk gas is likely to have been dissipated
47: in $\sim$ 3 Myr, well before such a core could accrete a significant gaseous
48: envelope (Bally et al. 1998; Bric\~eno et al. 2001; Haisch, Lada, \& Lada
49: 2001; Eisner \& Carpenter 2003).
50:
51: Microlensing surveys have also found evidence for two gas giant planets
52: with masses of $\sim 1.5 M_{Jup}$ orbiting M dwarfs (Bond et al. 2004;
53: Udalski et al. 2005). Because the microlensing signal produced by gas giant
54: planets is considerably stronger than that produced by super-Earths,
55: these four detections suggest that given the limited signal-to-noise
56: ratios of the ground-based photometry of microlensing events,
57: super-Earths must be signficantly more frequent companions to
58: M dwarfs than gas giants. Thus these four detections have been taken
59: as supportive of core accretion's inability to form gas giants around
60: low-mass stars (Beaulieu et al. 2006).
61:
62: Radial velocity surveys were the first to discover the existence
63: of gas giants and super-Earths around M dwarf stars. The M dwarf
64: Gl 876 has an outer pair of gas giant planets as well as an
65: inner super-Earth (Rivera et al. 2005). The M dwarfs Gl 436 and Gl 581
66: also appear to be orbited by super-Earths on short-period orbits
67: (Butler et al. 2004; Bonfils et al. 2005). M dwarf planet surveys
68: have only been underway for a few years, but they have already
69: revealed that the frequency of close-in gas giants around M dwarfs appears
70: to be lower than that around F, G, and K dwarfs (Endl et al. 2003, 2006).
71: With time, these surveys will determine the frequency of longer period
72: gas giants orbiting M dwarfs. Given the results to date of the radial
73: velocity and microlensing surveys, however, there is a clear need to
74: explain the formation of both gas giants and super-Earths around
75: M dwarf stars.
76:
77: We show here that disk instability can explain the formation
78: of the gas giant planets orbiting M dwarfs found by both microlensing
79: and radial velocity, as well as the super-Earths found by microlensing.
80: We begin by presenting more details about a disk instability
81: model (5CH) from Boss (2006a) that showed the possibility
82: of forming gas giant protoplanets about M dwarfs, and then develop
83: the reasoning behind the scenario for super-Earth formation
84: by photoevaporation of gaseous protoplanets orbiting M dwarfs.
85:
86: \section{Numerical Methods and Initial Conditions}
87:
88: Model 5CH (Boss 2006a) was calculated with a finite volume code
89: that solves the three dimensional equations of hydrodynamics and
90: radiative transfer, as well as the Poisson equation for the gravitational
91: potential. The code is second-order-accurate in both space and time
92: (Boss \& Myhill 1992) and has been used and discussed extensively in
93: previous disk instability studies (e.g., Boss 2003, 2005, 2006a).
94:
95: The equations are solved on a spherical coordinate grid with
96: $N_r = 101$, $N_\theta = 23$ in $\pi/2 \ge \theta \ge 0$,
97: and $N_\phi = 512$. The radial grid extends from
98: 4 AU to 20 AU with a uniform spacing of $\Delta r = 0.16$ AU.
99: The $\theta$ grid is compressed toward the midplane in order to ensure
100: adequate vertical resolution ($\Delta \theta = 0.3^o$ at the midplane).
101: The $\phi$ grid is uniformly spaced to prevent any azimuthal bias.
102: The central protostar wobbles in response to the growth of
103: disk nonaxisymmetry, preserving the location of the center
104: of mass of the star and disk system. The number of terms in the
105: spherical harmonic expansion for the gravitational potential of the disk
106: is $N_{Ylm} = 48$.
107:
108: Model 5CH calculated the evolution of a $0.5 M_\odot$
109: protostar surrounded by a protoplanetary disk with a mass of 0.065
110: $M_\odot$ between 4 AU and 20 AU. The initial protoplanetary
111: disk structure was based on an approximate vertical density
112: distribution (Boss 1993). The initial disk temperatures were
113: derived from the models of Boss (1995), with midplane temperatures of 300 K
114: at 4 AU, decreasing monotonically outward to a distance of $\sim 6.7$ AU,
115: where they were assumed to become uniform at an outer disk temperature
116: of $T_o = 50$ K. The outer disk was then initially marginally
117: gravitationally unstable in terms of the gravitational stability
118: parameter $Q$, with an initial minimum value of $Q = 1.5$.
119:
120: \section{Gaseous Protoplanet Formation}
121:
122: Figures 1 and 2 show that model 5CH formed a number of
123: clearly defined clumps after 215 years of disk evolution
124: around a protostar with a mass of $0.5 M_\odot$. The clumps
125: grow inside spiral arms, which form in the innermost, marginally
126: gravitationally unstable regions of the disk ($\sim$ 8 AU),
127: because of the combination of relatively low $Q$ and short
128: orbital periods there (see Figure 1 of Boss 2006a). This is
129: largely a result of the initial temperature profile assumed
130: for the disk midplane (Boss 1995). Such a profile is to be expected
131: for a protoplanetary disk in a region of low-mass star formation,
132: or in any star-forming region prior to the formation of the first
133: high-mass stars.
134:
135: Table 1 lists the maximum densities in the four clumps evident
136: in Figures 1 and 2, along with the clump masses $M_c$ in units of
137: the Jupiter mass $M_{Jup}$, the Jeans mass $M_J$ at the average
138: density and temperature of each clump, and the instantaneous
139: values of the orbital semimajor axis and orbital eccentricity of
140: each clump. In Figure 1, the first clump is located at
141: 11 o'clock, the second at 3 o'clock, the third at 7 o'clock,
142: and the fourth at 8 o'clock. Table 1 shows that each clump has a
143: mass well in excess of the local Jeans mass, showing that these
144: clumps are gravitationally bound. Their effective spherical
145: radii are comparable to the critical tidal radii at their
146: orbital distances, implying stability against tidal disruption
147: by the protostar's tidal forces as well.
148:
149: Because of the fixed nature of the numerical grid, the code
150: is not able to provide the locally enhanced spatial resolution
151: that these high density clumps require for their further
152: evolution to be calculated correctly. While clump densities
153: and lifetimes can be increased as the numerical spatial
154: resolution is increased (Boss 2005), with a fixed grid code
155: eventually the clumps are sheared apart. Meanwhile, new
156: clumps continue to form and take their place. Calculations where
157: the dense clumps are replaced by virtual protoplanets suggest
158: that the protoplanets should be able to orbit stably for
159: an indefinite period of time, even as the marginally gravitationally
160: unstable disk continues to transport mass inward to the
161: central protostar (Boss 2005). Similarly, SPH code calculations
162: with a locally-defined smoothing length by Mayer et al. (2002)
163: have shown that dense clumps should be able to survive their
164: subsequent orbital evolution, though mergers, scatterings, and
165: significant orbital evolution (both inward and outward) are to be
166: expected when multiple clumps form, as in model 5CH.
167:
168: While we cannot therefore predict the final outcome of model 5CH
169: with any degree of certainty, the model suggests that
170: one or more gas giant protoplanets with masses on the order of one
171: to a few Jupiter masses could form by disk instability around an M
172: dwarf star, with initial semimajor axes on the order of $\sim 10$ AU.
173: In addition to mutual scattering events, Type II migration during the
174: disk's lifetime could force gas giants to migrate closer to their
175: stars. It should be noted that model 5CH was not designed to
176: attempt to form clumps {\it in situ} at the $\sim$ 2.5 AU orbital
177: separation at which the microlensing techinique is most sensitive,
178: but that with minor changes in the assumed initial disk profiles
179: (i.e., a cooler inner disk), such an outcome would be likely.
180:
181: \section{Super-Earth Formation}
182:
183: If the M dwarf star and disk system represented by model 5CH
184: had formed in a region of low-mass star formation like Taurus
185: or Ophiuchus, the M dwarf would be expected to be accompanied by
186: one or more $\sim 1 M_{Jup}$ gas giant planets orbiting at
187: distances of $\sim$ 10 AU or less, possibly explaining the
188: two gas giant planet microlensing detections (Bond et al. 2004;
189: Udalski et al. 2005). However, most stars are formed in
190: regions of high-mass star formation (Lada \& Lada 2003),
191: similar to the Orion and Eta Carina nebulae, where protoplanetary
192: disks are subjected to a withering flux of FUV/EUV radiation from the
193: nearby O stars (e.g., Bally et al. 1998). In the Eta Carina nebula, FUV/EUV
194: fluxes are a factor of $\sim 100$ times higher than in Orion, yet
195: protoplanetary disks are as commonplace in Carina as in Orion
196: (Smith, Bally, \& Morse 2003). Armitage (2000) found that the EUV flux
197: alone in an Orion-like cluster was sufficient to photoevaporate
198: gaseous disks within $\sim$ 1 Myr around stars within 0.3 pc of the
199: massive stars. In larger clusters like Eta Carina, similarly
200: rapid photoevaporation would occur for disks within 3 pc of the O stars.
201:
202: Boss, Wetherill, \& Haghighipour (2002) suggested that the Solar System
203: was formed in a region of future high-mass star formation, such that after
204: the massive stars formed, their FUV/EUV radiation was able to photoevaporate
205: away not only the outer regions of the solar nebula, but also the gas
206: envelopes of the two outermost gas giant protoplanets formed
207: by disk instability (Boss 2003), stripping these two gaseous
208: protoplanets down to rock/ice cores with only minor gaseous envelopes,
209: i.e., turning them into the two ice giant planets, Uranus and Neptune.
210:
211: One critical component of this scenario for ice giant planet formation
212: is for the heavy elements to coagulate into dust grains and sediment
213: down to the centers of the protoplanets faster than the protoplanets
214: contract to planetary densities. Boss (1998) estimated that core
215: formation in gaseous protoplanets would occur in $\sim 10^3$ yr,
216: with a $1 M_{Jup}$ protoplanet being able to form at most a
217: $6 M_\oplus$ rock/ice core. Helled, Kovetz, \& Podolak (2006)
218: performed a more detailed analysis, and confirmed that dust grains
219: would settle down to form a central core in $\sim 10^3$ yr in
220: a non-convecting protoplanet. When the effects of convective
221: turbulence were included, the grains grew faster and reached
222: the core in $\sim 30$ yr. Helled, Podolak, \& Kovetz (2006)
223: found that a $1 M_{Jup}$ protoplanet requires $\sim 3 \times 10^5$ yr
224: to contract from a radius of $\sim 0.5$ AU to $\sim 0.1$ AU.
225: During this slow contraction phase, the protoplanet is able
226: to accrete a significant number of km-sized planetesimals
227: by gas drag capture in the protoplanet's outer layers. These
228: planetesimals will either be added to the solid core or will
229: remain in the protoplanet's envelope, but in either case the
230: protoplanet will be highly enriched in heavy elements compared
231: to the solar composition (Helled, Podolak, \& Kovetz 2006).
232: While much remains to be determined, it seems likely that
233: gas giants with core masses and envelope enrichments similar
234: to those of Jupiter and Saturn (Saumon \& Guillot 2004) can
235: be formed by disk instability.
236:
237: The key factor for whether a gaseous protoplanet becomes a gas giant
238: or an ice giant is the critical orbital radius $r_e$ outside of which
239: photoevaporation can remove the disk gas, and hence the protoplanetary
240: envelope gas. A gravitational radius $r_g$ can be defined to be the
241: orbital radius where the sound speed of the UV-heated gas equals the
242: gravitational escape speed from the protostar
243:
244: $$ r_g \approx {G M_p \over c_s^2},$$
245:
246: \noindent
247: where $M_p$ is the protostar mass and $c_s \approx 10$ km s$^{-1}$ for gas
248: heated by EUV radiation and $c_s \approx 3$ km s$^{-1}$ for FUV radiation.
249: Depending on the details of the photoevaporation model, $r_e$ is
250: expected to be smaller than $r_g$, with $r_e \sim 0.5 r_g$
251: (Johnstone et al. 1998) or even smaller (Adams et al. 2004).
252: With $r_e \sim 0.5 r_g$, for a G dwarf star like the Sun,
253: $r_e \sim 50$ AU for FUV radiation and $r_e \sim 5$ AU for EUV radiation.
254: Adams et al. (2004) argue that these values of $r_e$ could be
255: smaller by factors of as large as 5, i.e., to as small as $\sim 10$ AU
256: to $\sim 1$ AU for FUV and EUV, respectively. While the exact
257: value of $r_e$ for any protostar will depend on the relative
258: amounts of FUV and EUV received, we can calibrate $r_e$ by
259: noting that if photoevaporation was involved in the formation
260: of the Solar System's giant planets (Boss et al. 2002), then
261: evidently $r_e < 9$ AU, in order to result in Saturn's bulk
262: composition (assuming little orbital migration of Saturn after its
263: formation at $\sim$ 9 AU).
264:
265: These critical orbital radii depend linearly on the mass of the protostar.
266: For an M dwarf star with 1/3 the mass of the Sun, $r_e$ would
267: then be expected to be $\sim$ 3 AU or less, assuming the M dwarf protostar
268: was exposed to the same FUV/EUV environment as the solar nebula.
269: Gaseous protoplanets orbiting at this distance or beyond would lose the
270: bulk of their hydrogen gas by photoevaporation, leaving
271: a protoplanet composed primarily of the residual heavy elements
272: (helium gas would also be lost by entrainment in the photoevaporative
273: hydrogen flow). As a result, the final planet would be composed almost
274: exclusively of the core and envelope heavy elements.
275: The models of Helled, Podolak, \& Kovetz (2006) suggest that a
276: $\sim 1 M_{Jup}$ protoplanet formed by disk instability could
277: accrete as much as $\sim 30 M_\oplus$ of km-sized planetesimals
278: from the protoplanet's feeding zone, using the same assumptions
279: as are used in core accretion models (e.g., Pollack et al. 1996).
280: Planetesimal scattering was neglected in these models, so this
281: estimate appears to be a rough upper bound. [Cometesimal scattering
282: by the giant planets is thought to be the source of the Oort Cloud
283: comets and of the scattered disk component of the Kuiper Belt.]
284:
285: Model 5CH produced several $\sim 1 M_{Jup}$ gaseous protoplanets
286: orbiting an M dwarf at $\sim 10$ AU. If this system formed in
287: a region of future intense FUV/EUV radiation, these protoplanets would
288: be stripped down to cores composed of heavy elements, with
289: masses no larger than $\sim 30 M_\oplus$. Assuming that mutual
290: scattering and/or Type II migration had resulted in one of
291: these planets ending up on a $\sim 2.5$ AU orbit, it could then
292: be detected as a super-Earth by microlensing surveys.
293:
294: \section{Conclusions}
295:
296: Boss (2006b) pointed out that if the Orion-Carina scenario of
297: Boss et al. (2002) was responsible for the formation of the
298: Solar System's ice giant planets, then a possible test of
299: this combination of disk instability and photoevaporative losses
300: would be to see if the dividing line ($r_e$) between gas
301: giants and ice giants depends on the stellar mass: for lower
302: mass stars, this critical radius should decrease proportionately.
303: It remains for this prediction to be tested by future extrasolar
304: planet searches.
305:
306: Because most M dwarf stars are expected to have been subjected to
307: a high FUV/EUV radiation environment (Lada \& Lada 2003),
308: most M dwarf planets found by microlensing surveys would
309: be expected to be super-Earths rather than gas giants, as
310: seems to be the case so far, albeit based on a small sample of
311: only four detections to date. Future microlensing detections
312: will be important to determine if this provisional interpretation
313: is correct.
314:
315: The lower frequency of gas giant planets on short-period orbits
316: around M dwarfs compared to F, G, and K dwarfs (Endl et al. 2003,
317: 2006) may be a result of faster inward orbital migration around
318: the more massive dwarfs. The well-known correlation of the presence
319: of gas giants with the metallicity of the host star appears to
320: be strongest for short-period planets (Sozzetti 2004), consistent
321: with the expectation that Type II inward migration will be faster
322: in metal-rich disks (Livio \& Pringle 2003) and could thus result in a
323: higher frequency of short-period gas giants (Boss 2005). The rate of
324: Type II migration depends on the disk's kinematic viscosity $\nu$, and in
325: standard viscous accretion disk theory (e.g., Ruden \& Pollack 1991)
326: $\nu = \alpha c_s h$, where $\alpha$ is a free parameter, $c_s$ is
327: the sound speed, and $h$ is the disk thickness. M dwarfs will have
328: cooler disks than G dwarfs (Boss 1995) because of their shallower
329: gravitational potential wells and their (presumed) proportionately
330: lower disk masses and hence smaller optical depths. As a result,
331: M dwarf disks should be thinner and have smaller sound speeds than
332: G dwarf disks, leading to smaller values of $\nu$ and hence longer
333: Type II migration times. M dwarf planets should thus be less
334: likely to undergo significant Type II migration prior to removal
335: of their disk gas.
336:
337: There is still an important role to play for the first step of
338: the core accretion process in forming planets around M dwarfs.
339: One of the short-period super-Earths found by the radial velocity
340: surveys of M dwarfs (Gl 876 -- Rivera et al. 2005) is known to be
341: accompanied by two outer gas giant planets, implying that the super-Earth
342: formed interior to the gas giants. The scenario presented in this
343: paper would not be able to explain the formation of the Gl 876 system,
344: unless the planets were able to interchange their radial ordering,
345: which seems unlikely. Hence the Gl 876 super-Earth is likely to have been
346: formed interior to its gas giants by the same collisional accumulation
347: process that led to the formation of the terrestrial planets in our Solar
348: System (e.g., Wetherill 1996). Taken as a whole, M dwarfs thus appear to
349: present strong evidence for the formation of the same three classes
350: of planets found in our Solar System: inner terrestrial planets
351: formed by collisional accumulation, and outer gas giants or rock/ice
352: giants (super-Earths) formed by disk instability, either in
353: the absence of, or in the presence of, strong fluxes of FUV/EUV
354: radiation, respectively.
355:
356: I thank Michael Endl for discussions about M dwarf planets,
357: the referee for several improvements to the manuscript,
358: and Sandy Keiser for her computer systems expertise. This research
359: was supported in part by NASA Planetary Geology and Geophysics
360: grant NNG05GH30G and by NASA Astrobiology Institute grant NCC2-1056.
361: The calculations were performed on the Carnegie Alpha Cluster,
362: the purchase of which was partially supported by NSF Major Research
363: Instrumentation grant MRI-9976645.
364:
365: \clearpage
366:
367: \begin{references}
368:
369: \reference{r}
370: Adams, F. C., et al. 2004, ApJ, 611, 360
371:
372: \reference{r}
373: Armitage, P. J. 2000, A\&A, 362, 968
374:
375: \reference{r}
376: Bally, J., et al. 1998, AJ, 116, 854
377:
378: \reference{r}
379: Beaulieu, J.-P., et al. 2006, Nature, 439, 437
380:
381: \reference{r}
382: Bond, I. A., et al. 2004, ApJ, 606, L155
383:
384: \reference{r}
385: Bonfils, X., et al. 2005, A\&A, 443, L15
386:
387: \reference{r}
388: Boss, A. P. 1993, ApJ, 417, 351
389:
390: \reference{r}
391: ------. 1995, Science, 267, 360
392:
393: \reference{r}
394: ------. 1998, ApJ, 503, 923
395:
396: \reference{r}
397: ------. 2003, ApJ, 599, 577
398:
399: \reference{r}
400: ------. 2005, ApJ, 629, 535
401:
402: \reference{r}
403: ------. 2006a, ApJ, in press (20 May 2006 issue)
404:
405: \reference{r}
406: ------. 2006b, in Planet Formation 2004: Observations, Experiments,
407: and Theory, H. Klahr \& W. Brandner, eds., Cambridge: Cambridge
408: University Press, in press
409:
410: \reference{r}
411: Boss, A. P., \& Myhill, E. A. 1992, ApJS, 83, 311
412:
413: \reference{r}
414: Boss, A. P., Wetherill, G. W., \& Haghighipour, N. 2002, Icarus, 156, 291
415:
416: \reference{r}
417: Brice\~no, C., et al. 2001, Science, 291, 93
418:
419: \reference{r}
420: Butler, R. P., et al. 2004, ApJ, 617, 580
421:
422: \reference{r}
423: Eisner, J. A., \& Carpenter, J. M. 2003, ApJ, 598, 1341
424:
425: \reference{r}
426: Endl, M., et al. 2003, AJ, 126, 3099
427:
428: \reference{r}
429: ------. 2006, ApJ, submitted
430:
431: \reference{r}
432: Gould, A., et al. 2006, ApJ, submitted
433:
434: \reference{r}
435: Haisch, K. E., Lada, E. A., \& Lada, C. J. 2001, ApJ, 553, L153
436:
437: \reference{r}
438: Helled, R., Kovetz, A., \& Podolak, M. 2006, in preparation
439:
440: \reference{r}
441: Helled, R., Podolak, M., \& Kovetz, A. 2006, Icarus, submitted
442:
443: \reference{r}
444: Johnstone, D., Hollenbach, D., \& Bally, J. 1998, ApJ, 499, 758
445:
446: \reference{r}
447: Lada, C. J., \& Lada, E. A. 2003, ARAA, 41, 57
448:
449: \reference{r}
450: Laughlin, G., Bodenheimer, P., \& Adams, F. C. 2004, ApJ, 612, L73
451:
452: \reference{r}
453: Livio, M., \& Pringle, J. E. 2003, MNRAS, 346, L42
454:
455: \reference{r}
456: Mayer, L., et al. 2002, Science, 298, 1756
457:
458: \reference{r}
459: Pollack, J. B., et al. 1996, Icarus, 124, 62
460:
461: \reference{r}
462: Rivera, E., et al. 2005, ApJ, 634, 625
463:
464: \reference{r}
465: Ruden, S. P., \& Pollack, J. B. 1991, ApJ, 375, 740
466:
467: \reference{r}
468: Saumon, D., \& Guillot, T. 2004, ApJ, 609, 1170
469:
470: \reference{r}
471: Smith, N., Bally, J., \& Morse, J. A. 2003, ApJ, 587, L105
472:
473: \reference{r}
474: Sozzetti, A. 2004, MNRAS, 354, 1194
475:
476: \reference{r}
477: Udalski, A., et al. 2005, ApJ, 628, L109
478:
479: \reference{r}
480: Wetherill, G. W. 1996, Icarus, 119, 219
481:
482: \end{references}
483:
484: \clearpage
485: \begin{deluxetable}{cccccc}
486: \tablecaption{Clump properties for model 5CH at 215 yr. \label{tbl-1}}
487: %\footnotesize
488: %\tablewidth{0pt}
489: \tablehead{\colhead{Clump} &
490: \colhead{$\rho_{max}$ (g cm$^{-3}$)} &
491: \colhead{$M_c/M_{Jup}$} &
492: \colhead{$M_J/M_{Jup}$} &
493: \colhead{$a$ (AU)} &
494: \colhead{$e$} }
495: \startdata
496: 1 & $6.2 \times 10^{-10}$ & 1.0 & 0.74 & 8.7 & 0.040 \nl
497: 2 & $2.7 \times 10^{-9}$ & 1.1 & 0.51 & 7.9 & 0.10 \nl
498: 3 & $1.4 \times 10^{-9}$ & 1.3 & 0.57 & 9.0 & 0.093 \nl
499: 4 & $1.6 \times 10^{-9}$ & 1.4 & 0.72 & 8.0 & 0.011 \nl
500: \enddata
501: \end{deluxetable}
502: \clearpage
503:
504:
505: \begin{figure}
506: \vspace{-2.0in}
507: \plotone{f1.eps}
508: \caption{Equatorial density contours for model 5CH after 215 yrs of
509: evolution. The entire disk is shown, with an outer radius of 20 AU and an
510: inner radius of 4 AU, through which mass accretes onto the central protostar.
511: Hashed regions denote spiral arms and clumps with densities higher than
512: $10^{-10}$ g cm$^{-3}$. Density contours represent factors of two change
513: in density. The four clumps described in Table 1 are numbered sequentially
514: in counterclockwise order as they appear in this Figure, starting with
515: clump \#1 at 11 o'clock.}
516: \end{figure}
517:
518: \begin{figure}
519: \vspace{-2.0in}
520: \plotone{f2.eps}
521: \caption{Same as Figure 1, but with cross-hatching removed to
522: reveal the structure of the density contours in the densest regions.}
523: \end{figure}
524:
525: \end{document}
526:
527: