; peeses) ' wy Bite att y ir anai er ane: ree! URS mea Raat f . Nee y bet ! iba hit sabres te ak Tes tapi tes inh Biriatt bea r a i) PERO M, We setae th : ty usaen rev tat seule i Seth! het REAR: Dantes ai ties iin ned oust Pee CU 4 astern iespsiat yates yee pi jeter ( Diahtd ~ a i petit A re ettiateen beta pit tee i i) Hi th th ye FFE “s revert bs ni + hy +2 et 32 2; ae oe e525) hescouscoss Shr Seee Qe, rve ra? borer e te e Lee UNIVERSITY OF CALIFORNIA BUBLICATIONS ee eh AS 2) oe IN. ZOOLOGY Vol. 12, No. 4, pp. 51-294, pls. 3-13, 9 text figs. March 20, 1914 AN ACCOUNT OF THE MAMMALS AND BIRDS OF THE LOWER COLORADO VALLEY WITH ESPECIAL REFERENCE TO THE DISTRIBU- ~ _ TIONAL PROBLEMS PRESENTED BY JOSEPH GRINNELL UNIVERSITY OF CALIFORNIA PRESS BERKELEY sear-aeen UNIVERSITY OF CALIFORNIA PUBLICATIONS as Note.—The University of California Publications are offered in exchange for the publi- cations of learned societies and institutions, universities and libraries. Complete lists of all the publications of the University will be sent upon request. For sample copies, lists of publications or other information, address the Manager of the University Press, Berkelsy, Oalifornia, U. 8. A. All matter sent in exchange should be addressed to The Exchange Department, University Library, Berkeley, California, U. S. A. } { OTTO HARRASSOWITZ, BR. FRIEDLAENDER & SOHN, { LEIPZIG. BERLIN. : he Agent for the series in American Arch- Agent for the series in American Arch- aeology and Ethnology, Classical Philology, aedlogy and Ethnology, Botany, Geology, Education, Modern Philology, Philosophy, Geography, Mathematics, Pathology, Physi- - Psychology, History. ology, Zoology, and. Memoirs, \ ZOOLOGY.—W. E. Ritter and C. A. Kofoid, Editors. Price per volume, $3.50; beginning’ with vol. 11, $5.00. This series contains the contributions from the Department of Zoology, from the Marine Laboratory of the Scripps Institution for Biological Research, at La Jolla, California, and from the California Museum of Vertebrate Zoology in Berkeley. Cited as Univ. Calif. Publ. Zool. Volume 1, 1902-1905, 317 pages, with 28 plates’ 2... lee cece ceacenectceenecensecee $3.50 Volume 2 (Contributions from the Laboratory of the Marine Biological Associa- tion of San Diego), 1904-1906, xvii + 382 pages, with 19 plates... $3.56 Volume 3, 1906-1907, 383 pages, with 23 plates oe ..c.cccceccecececseccccenecstecescesenpeceseeceeesseae $3.50 Volume 4, 1907-1908, 400 pages, with 24 plates ..... Volume 5, 1908-1910, 440 pages, with 34 plates Volume 6, 1908-1911, 478 pages, with 48 plates Vol. 7. (Contributions from the Museum of Vertebrate Zoology.) 1. Two New Owls from Arizona, with Description of the J: ayoual. Plum- age of Strix occidentalis occidentalis (Kantus), by Harry 8S. Swarth. Pp. 1-8. May, 1910 2. Birds and Mammals of the 1909 Alexander Alaska Expe Harry S. Swarth. Pp. 9-172; plates 1-6; 3 text-figures. January, 1911. 1.50 8. An Apparent Hybrid in the Genus Dendroica, by Walter P. Taylor. EPs LSAT. FO PLUAry ys LOU er a a aa 05 4. The Linnet of the Hawaiian Islands: a Problem in Speciation, by Joseph Grinnell, Pp. 179-195. February, 1911 220.000 15 5. The Modesto Song Sparrow, by Joseph Grinnell. Pp. 197-199. - Feb- TURK Yy LO oe ie eee SE Se eS cease fuse .05 6. Two New Species of Marmots from Northwestern America, by H. 8. Swarth.- Pp. 201-204. February, 19V4 oo. c.c2i eel ccicc cece ceeeeccnnsececnseess 06 7. Mammals of the Alexander Nevada Expedition of 1909, by Walter P, Taylor. bp. 205-3074) ine, PORT ro ee aa tea a eat 1.00 8. Description of a New Spotted Towhee from the Great Basin, by J. Grinnell, Pp. 309-311. August, LOUD oo... cic ciceleceee cieceeedencecanecoeeee .05 9. Description of a New Hairy Woodpecker from Southeastern Alaska, by H. S. Swarth. Pp. 318-318... October, 1914: ...2--.2ci ences ceed ane 05 10. Field Notes on Amphibians, Reptiles and Birds of Northern Humboldt County, Nevada, with a Discussion of Some of the Faunal Features of the Region, by Walter P. Taylor. Pp. 319-486, plates 7-12. FF ODIUATY, 194 Fens ee rn eshe eeetnd ug Ses ancaegae take 1.00 Index, pp. 437-446, Vol. 8 1, The Vertical Distribution of Hucalanus elongatus in the San Diego Region during 1909, by Calvin O. Esterly. Pp. 1-7. May, 1911 ....... 10 2. New and Bare Fishes from Southern California, by Edwin Chapin Starks and William M. Mann. Pp. 9-19, 2 text-figures. July, 1911. ..10 8. Classification and Vertical Distribution of the Chaetognatha of the San Diego Region, Including Redescriptions of Some Doubtful Species of the Group, by Ellis L. Michael. Pp. 21-186, pls. 1-8: December, 1911. 1.75 4, Dinoflagellata of the San Diego Region, IV. The Genus Gonyaulaz, with Notes:on Its Skeletal Morphology and a Discussion of. Its Generic and Specific Characters, by Charles Atwood Kofoid. Pp. 187-286, plates 9-17. UNIVERSITY OF CALIFORNIA PUBLICATIONS IN ZOOLOGY Vol. 12, No. 4, pp. 51-294, pls. 3-13, 9 text figs. March 20, 1914 AN ACCOUNT OF THE MAMMALS AND BIRDS OF THE LOWER COLORADO VALLEY WITH ESPECIAL REFERENCE TO THE DISTRIBUTIONAL PROBLEMS PRESENTED BY i r JOSEPH GRINNELL . (Contribution from the Museum of Vertebrate Zoology of the University of California) CONTENTS PAGE TETAS ROOT, ee eee ee Ree 51 AT GUI O TAY Y iy tenets cee cese woes teessepeeocest sess 53 Desenipionmote thex Colorado: VRUVCN, j.2c-cececscctcce ce sece noses wren sec en ah wean aee ese Sek lose aesetnnctseese 57 Aonalvand=faunaleposition Of Chey Teen -o:. ce. 2e ota aac cen ce neccensnreececerneeeeereevorcecase 62 NRETASSOCIALIONAlaAT eas OLsuMe MOP LOM cetera sasse a seee si asa sacaccstcnsaneacsnesnossaeses ze. 66 The Colorado River asa highway of dispersal ana center of differentiation of species ...........--. os 97 The Colorado River as a hindrance to the dispersal of species - - 100 The problem of barriers with regard to birds and mammals - 107 @heckelistro fe ven bins) ee cece esse acres cceaee see cans Sect sees oet cer ocee cb ct cortves octenet eee ac caleaceetase 110 General accounts of the birds: local Geranationt migration, variation, molt, 1 EE ORS ee sae eee Pe Se ee ae Se en eRe 113 (CINCY SI EL oop UT AES 000 eae 217 General accounts of the mammals: local ceaenod variation, habits ..... 218 GONE CY CULO oe ceen scorn ves cen csc ee ee a cncee ce tae pena c hesetesassnec con ceiueecuswesocstctooorvccsetsaenastatavean 269 INTRODUCTION The southeastern frontier of California lies in the heart of a vast desert region possessing a fauna and flora of notable variety and peculiar specialization of forms. Cutting through the center of this desert area flows the great Colorado River. Politically, this river 52 University of California Publications in Zoology | Vou. 12 marks the southeastern boundary of California, comprising the interval between the Nevada line and the Mexican line, and separates our state from Arizona. Biogeographically, the Colorado River completely bisects the desert area through which it flows, the resultant two di- visions with their xerophilous fauna and flora being separated, not only by the stream itself, but also by the riparian strips of more or less width which flank the river immediately on either side. Two objects were in mind to justify the selection of the Colorado valley for extended field-work in vertebrate zoology: (1) The literature pertaining to the birds and mammals of the region was fragmentary, relatively little work having been published since the early and incom- plete reports of Woodhouse (1853), Cooper (1861, 1868, 1869), and Coues (1866). An extended knowledge of the composition of the vertebrate fauna of the southeastern frontier of California was a desideratum. (2) The effect of the Colorado River, with its riparian strips sharply contrasted against the contiguous desert areas, upon the distribution of the animals concerned, promised to provide data of importance relative to the general problem of barriers. Miss Annie M. Alexander, founder of the California Museum of Vertebrate Zoology, not only warmly approved of the plan to under- take field-work along the lower Colorado River with the above objects in view, but generously provided the necessary cost of the expedition in addition to her regular appropriation for the support of the museum. The plan was carried out in the three months of 1910 from February 15 to May 15, inclusive. The writer took part in the field work in person throughout the entire time, and was assisted by Mr. Frank Stephens, Mr. Joseph Dixon, and Mr. L. Hollister Jones. The party began work at Needles, and proceeded by boat from place to place down the river, the last station being on the California side below Yuma and close to the Mexican line. The locations of the various collecting stations established are given in the itinerary. The three months’ field-work yielded 1,272 specimens of mammals, 1,374 birds, 443 reptiles and amphibians, 22 sets of birds’ eggs and nests, a few fishes, and a collection of the more conspicuous plants. All of this material now forms part of the collections of the Cali- fornia Museum of Vertebrate Zoology, except the plants, which are deposited in the Herbarium of the Department of Botany of the University of California, and the few fishes, which are added to the ichthyological collection in the Zoological Department of Leland Stan- ford Junior University. 1914] Grinnell: Mammals and Birds of the Colorado Valley 53 The present paper concerns itself with the mammals and birds of the region, and is based upon the collections and field-notes obtained in 1910, as above enumerated. No attempt has been made as yet to work up the reptiles and amphibians, though these promise interesting results. The writer’s efforts have been concentrated upon the mam- mals and birds, with a view to establishing first of all the systematic status of the various included forms; in other words, to determine their relationships with similar species or races of the surrounding regions. Such determination has been performed in a manner wholly satisfactory to the writer in only a part of the critical cases. As must always obtain in work of this sort, lack of material in certain direc- tions often puts abrupt limitations upon investigation. The obvious principle has been followed, that no generalized treat- ment of a conerete subject like geographic distribution is justifiable upon any but the soundest basis of systematic analysis. Because of the close dependence of most mammals and birds upon plants, the latter have an important place in any treatment of animal distribution. The set of plants secured by the Colorado River Expedi- tion has been named by Professor H. M. Hall, of the University of California. Names so provided are used throughout the present paper, particularly in the discussion of associational areas. My sincere acknowledgements are hereby extended. to Professor Charles H. Gilbert, Department of Zoology, Leland Stanford Junior University, and to Professor Charles A. Kofoid, Department of Zoology, University of California, for critical suggestions concern- ing general considerations; to Professor H. M. Hall, of the University of California, for critical reading of the chapter on associations; and to Messrs. Harry S. Swarth and Walter P. Taylor, fellow staff- members in the California Museum of Vertebrate Zoology, for help from time to time on many points of detail. ITINERARY Our party assembled at the town of Needles, California, on the evening of February 14, 1910, and on the following day our first collecting station was established in the river bottom close by. A scow was constructed for the transportation of our rather bulky outfit, while the skiff purchased, being a much readier means of loco- motion, enabled us to traverse the river expeditiously where need be. It was found practicable to divide the party during a portion of the 54 University of California Publications in Zoology | Vou. 12 time, two of the men traveling separately with the skiff and a light outfit. All members of the party were working from the same base, except where indicated otherwise. The current was depended upon almost altogether to carry our boats from station to station. The following enumeration shows the location of the various sta- tions established from the initial point down the river, the names by which they are designated throughout this report, and the time spent at each. For their location reference should be made to the map (pl. 3). 1. Needles, California side; camp in river bottom one-fourth mile east of the railroad station; February 15 to 18. 2. Five miles below Needles, California side; camp in river bottom in an air line probably shehtly less than four miles south-southeast of Needles; February 18 to 23. In this and some other cases, even where a government map was at hand, the exact location on the flood bottom was indeterminable on the map, because of shifting of the river channel; for it frequently happens that in a single period of high water the topography of the riparian area is profoundly metamor- phosed. 3. Mellen, Arizona side; camp on narrow mud-bar within one-fourth mile north of the railroad station locally called Topock (==Mellen, on the Needles Special Map, U. S. G. 8S.) ; February 23 to March 1. The Santa Fe railroad crosses the Colorado River at this point. 4. Opposite The Needles, California side; camp on high bank about one mile northwest of ‘‘B. M. 465’’ of Needles Special Map, U.S. G.S8.; March 1 to 8. ‘‘The Needles’? are a group of sharp-peaked hills chiefly on the Arizona side of the river, and must be remembered as ce a totally distinet loeality from ‘‘Needles,’’ the name of the railroad town fifteen miles or more to the northwest. 5. Foot of The Needles, Arizona side; this station was worked from the camp on the opposite side of the river. Certain members of the party rowed back and forth morning and evening to tend trap lines; March 4 to 7. 6. Chemehuevis Valley, California side; camp in river bottom at lower end of this valley and close to base of hills, probably near “*B. M. 438”’ of Parker quadrangle, U.S. G. S.; March 8 to 11. With this and the succeeding two stations uncertainty exists as to exact location of our base camps. The maps we had at that time were inaccurate. The Parker quadrangle, U. S. G. S., was not issued until OTe 1914] Grinnell: Mammals and Birds of the Colorado Valley 55 7. Lower Chemehuevis Valley, California side; camp at edge of whirlpool and in mouth of wash emanating from low hills, probably of Parker quadrangle, 7 about one mile west-northwest of ‘‘B. M. 418 U.S. G. S.; March 11 to 12. 8. Above Bill Williams River, Arizona side; camp in river bottom, probably within two miles above ‘Steamboat Rock’’ of Parker quad- rangle, U. S. G. S., and hence about eleven miles northwest of the mouth of [Bill] Wiliams River; March 12 to 15. 9. Parker, Arizona side; camp at the new Santa Fe bridge; this was merely an over-night stop, and very little collecting was done; Mareh 15 to 16. 10. Riverside Mountain, California side; camp at mouth of wash, close to ‘‘B. M. 405”? of Parker quadrangle, U.S. G. S.; March 16 to 22. 11, 12. Above Blythe, California side. The party separated at Riverside Mountain, Stephens and Jones working along slowly on the lookout for beaver sign and making at least two over-night camps on the way to Ehrenberg. One of these stops, and the one by Grinnell and Dixon, were located some distance apart, in the river bottom adjacent to the mesa on the California side, and within six or eight miles above Blythe. March 22 to 24. 13. Ehrenberg, Arizona side; camp in river bottom within one- fourth mile below the town; March 24 to 30. 14, 15, 16. Below Ehrenberg, Arizona side, one station, and near Palo Verde, California side, two stations. The party again separated, and base camps established at three different points in the river bot- tom, estimated to be from ten to twenty-five miles below Ehrenberg. The changed course of the river made dependence upon maps impos- sible. March 28 to April 3. 17, 18. Opposite Cibola, California side, two stations ; main camp on high bank where river swung against mesa; somewhat above, that is, to the northwest of, rather than directly opposite, Cibola; Stephens and Jones worked from a second station about five miles below the main camp, April 5 to 6. March 31 to April 6. 19. Ten miles below Cibola, Arizona side; camp on narrow terrace between river and mesa, about a mile below an adobe ruin; April 6 to 10. 20. Twenty miles above Picacho, California side; camp in river bottom opposite Lighthouse Rock, and about a mile below the Draper ranch. In an air line this station was probably somewhat less than twenty miles north of Picacho. April 10 to 17. On fon) University of California Publications in Zoology — {| Vou. 12 21. Hight miles east of Picacho, California side; camp in river bottom at lower end of ‘‘Charlie’s Valley,’’ which in turn is just below Canebrake Canon. Because of the eastward swing of the river in the vicinity of Picacho, eight miles east of Picacho is also eight miles below Picacho. April 17 to 21. 22. Five miles above (or north of) Laguna, Arizona side; camp in river bottom near base of first hills above silted-in area; April 21 to 26. 23. Four miles above (or north of) Potholes, California side ; visited April 23, from our camp on the opposite side of the river. 24. Potholes, California side; camp in river bottom just below the head-gate; April 26 to 29. 25. Four miles below Potholes, California side; camp in river bot- tom; Grinnell and Dixon; April 29 to May 2. 26. Five miles northeast of Ywma, California side; camp in river bottom; Stephens and Jones; April 30 to May 3. 27. Five miles above (northeast of ) Yuma, California side; camp in river bottom; Grinnell and Dixon; May 2 to 5. This station was but a half-mile or so from the preceding. 28. Yuma, Arizona side; camp about half a mile up the Gila River from its mouth, on the shore towards Yuma and within two miles east of that town; Stephens and Jones; May 3 to 7. 29. Near Pilot Knob, California side; camp on river bank at site of the old Hanlon Ranch, marked by a group of date palms. This is also the site of the ‘‘ American Girl Pump,’’ of the Yuma quadrangle, U.S. G.S.; but only rusting machinery and pipes, and a pile of ashes, marked the place at the time of our visit. Grinnell and Dixon; May 5 to 15; Stephens and Jones, May 7 to 15. The season’s field-work was concluded on May 15. As will have been noted, our collecting stations were usually estab- lished on alternate sides of the river successively. It proved impos- sible to propel the heavy scow directly across because of the strong current. In two places, however, a limited amount of collecting was done on exactly opposite sides of the river, this being accomplished through the use of the skiff alone. From each of the base camps as a center, hunting and trapping were carried on as far radially as proved practicable. It was. our effort to test every sort of ground available, that is, each association represented. Three of the four members in the party were constantly engaged in trapping for mammals. Trap lines were run in different directions, usually from the river’s edge back to the highest parts of 1914] Grinnell: Mammals and Birds of the Colorado Valley 57 the desert within reach, generally not farther than two miles from camp. It was, of course, possible to carry a single day’s hunt much farther back on to the desert. A distance of from four to seven miles from the river was occasionally reached in day’s tramps, for instance, at Riverside Mountain, twenty miles above Picacho, and at Pilot Knob. In the nature of the ease, our camps were always pitehed on the river bank where convenient landings could be made and the boats tied up safely. As is to be expected, the writer now regrets that extended work was not done at certain points searcely or not at all touched in our 1910 exploration. It would have been a desirable thing to work intensively on directly opposite sides of the river, where it eut through high ground, at three or more points; for example, at The Needles, below Cibola, and at the Laguna Dam. It must be remembered, however, that we had practically no information to start with, either as to the fauna or local topography along most of the route. We had it all to learn by our own efforts; and only as experience accumulated could the contrasting sets of facts come into relief, thereby directing our enquiries into more productive channels. DESCRIPTION OF THE COLORADO RIVER The portion of the Colorado River explored by the 1910 expedition is altogether below the famed Grand Cafon. There is no true ‘“cafion’’ below the Nevada line, the river flowing through its chan- nel in relatively sober fashion. Yuma is very close to 150 miles in a direct line due south of Needles. While following a general north- to-south course, the many deflections of the river, and especially its meandering through the big valleys, make the distance from Needles to Yuma by the way of the river approximate 285 miles. The altitude of the river at Needles is 460 feet, at Yuma, 125 feet, so that the average fall in this portion of the river is only a little more than one foot per mile. Two features of the Colorado River are extraordinary, when com- parison is made with other streams of the North American continent. (1) The amount of sediment always carried in suspension is very great, so that the flowing waters are quite opaque at all seasons of 58 University of California Publications in Zoology [ Vou. 12 the year. (2) There is an enormous increase in the volume of the river at the annual period of high water (see Newell, 1904, pp. 123— 161). The sediment load varies from one-tenth to two and one-tenth per cent, by weight, according to the rate of the current. The latter varies from a maximum of two and one-half miles per hour at low water up to even seven miles per hour at flood time in parts of the channel where conditions are most favorable. While it is probable that the Needles-to-Yuma section of the Colorado River has about reached erade, the extraordinary supply of sediment always fed from above and the ever fluctuating volume of water, result in a continued pro- cess of deposition and erosion, going on simultaneously in different parts of the river’s channel. At low water deposition exceeds erosion, but at flood time the increased transporting power of the river results in removal of large masses of soil from varying parts of the river bottom. The flow of the river varies from 4,000 to 100,000 cubic feet per second. The time of lowest water is in midwinter, that of highest flood, in June, at the time of melting snow among the sources of the Colorado River, in the Rocky Mountains. The period of high water is of short duration, about May 15 to July 1, while throughout the year fluctuations of less extent are liable to occur at any time. The river’s course is of two markedly different types: where it flows between rock walls among hills the channel is of fixed location and usually deep and narrow; in the large valleys the river may be broad and shallow, but at any rate of constantly shifting channel. At The Needles the river passes between high hills and the narrow chan- nel thus formed has two sharp bends in it. Such a gorge is locally called a ‘‘box cafion,’’ and at certain stages of the river is not without danger to small boats because of whirlpools and jutting reefs of rock. At low water the voyager is astonished at the elevation of the highest water marks on the rock walls above the surface of the river at the low stage. There is good reason for believing that the river is now aggrading its channel considerably above the apex of the delta where, of course, this process is certainly in progress. At any rate, sedimentation has formed broad flood-bottoms at intervals along the river’s course. These, large or small, are very similar in physical and floral features. The most extensive one of these flood-bottoms is sometimes known as the ‘‘Great Valley of the Colorado.’’ The settlements of Blythe, W914} Grinnell: Mammals and Birds of the Colorado Valley 59 Ehrenberg, Palo Verde and Cibola are located in it, and considerable activity has of late been directed towards farming the rich bottom lands. But the capriciousness of the river has proven a hindrance to marked development along this line. Shortly below the point where the river emerges from its rock-con- fined channel into the broadening valley, meandering begins. The detours of the stream increase in extent towards the lower end of the valley, the channel swinging from side to side in great curves, marked by minor curves along these courses, so that a sort of periodicity of meandering is exhibited. The most notable phenomenon in this con- nection is the progressive movement of these loops down the valley. The result is that in a short period of years, the major portion of the river’s flood-bottom is worked over in the path of this irresistible and continual shifting of the channel. The effect on the flora is obvious. Only in curves of the valley sheltered by abutting hills are trees given a chance to reach advanced age. The only trees capable of thriving on the unstable portion of the flood-bottom are such as grow rapidly, willows and cottonwood. As the observer floats along the main channel, winding through the ereat valley, especially during rising water, the destructive process in question is clearly in evidence. On the outside of each curve the river rapidly undercuts the bank, dislodging great masses of the fine sedimentary soil, itself laid down but a few years previously. The overgrowth of comparatively aged willow and cottonwood topples into the stream and is swept away by the swift current. Hundreds of acres may be thus appropriated by the river in a few days and within a short distance. On the other hand, on the inside of each curve, where the current is slow, and especially during falling stages of the water, the river is rapidly depositing sediment, in other words building up its bank towards the general level of the bottom lands. At the annually recurring periods of high water, when the entire flood bottom is inundated, layers of silt are deposited over the whole, thus tending to establish a uniform level. The observer, from any appropriate hill-top overlooking the valley, ean readily discern the regularly graded heights of tree growth which mark the successive ages of the land on which they grow. The year- old seedlings but a few inches in height form a erescent-shaped belt along the inside of each curve of the river, facing down the valley. Paralleling this and next in position back from the river is dense 60 University of California Publications in Zoology [ Vou. 12 two-year-old growth, succeeding which is a stand of still older growth. Because of the progressive trend of the process it is as a rule the oldest growth which becomes subject to the razing action of the river as its loops travel down the valley. Oceasionally cut-offs occur, thus interrupting the symmetry of behavior, and the river quickly establishes a new channel, portions of the old being left as lagoons, though these are usually short-lived because of the rapid sedimentation at recurring times of general over- flow. The bottom land immediately adjacent to the channel, where the latter has been fixed for some time, is usually higher than the lateral tracts. The depressions on one or the other side of the flood- bottom are generally drained at the lower end of the valley by a series of sloughs emptying into the main channel just above the constriction of the valley where the hills converge. At high water these lateral depressions are submerged to a depth of as much as twelve feet, as shown by actual measurement of the upper limit of the mud marks on the tree trunks. Whatever the water touches is discolored by a coat of fine sediment, and one thereby gains at any time of the year an accurate estimate of the depth of inundation in any part of the flood bottom at the previous period of high water (see sectional profiles, figs. A and B). In some places the overflow depressions have no drainage outlets. After times of flood, the water in them disappears by evaporation, which proceeds at a very rapid rate in this excessively arid and hot region. Areas of alkaline deposits are left, often stretches of bare, baked hard-pan surrounded by concentric belts of halophilous vege- tation. At rather infrequent intervals along the large valleys the river, in swinging, cuts beyond its usual limit and assaults the desert mesa. The process tends to add to the area of the regular flood plain and conduees to the formation of lateral bluffs. It is noteworthy that this aggressive work of the river is much more conspicuous on the west side of each valley than on the east side. The law of westward cutting of north-and-south flowing streams in the northern hemis- phere, as brought about by the earth’s rotation, is thus clearly illus- trated in the Needles-to-Yuma section of the Colorado River. The local use of the terms ‘‘first bottom’’ and ‘‘second bottom’’ in the large valleys is worth adopting in general discussions relative to ranges of animals. The first bottom is the portion of a valley subject to regular overflow, and is clearly marked by the presence of 1914] Grinnell: Mammals and Birds of the Colorado Valley 61 willows and cottonwood; the second bottom is of sufficiently higher elevation to be chiefly above high-water mark, and is characterized by mesquite, salt-bush, and rank clumps of creosote bush. The second bottom may be altogether wanting, or it may constitute a broad bench- like tract. The surface is often modified by alluvial deposits at the mouths of washes leading down from the adjacent desert, and by wind-blown sands which heap up about bushes, especially along the southeastern borders of the valleys. The Laguna Dam has had a pronounced modifying influence on the flora and fauna of the vicinity. The dam was built to a height of twelve feet above the mean level of the river at that point at the time of beginning construction. As soon as it was completed (in 1909) the retarded waters above began to deposit silt, and by May, 1910, the valley above had been silted in to a depth determined by the top of the dam. The water level had been raised conspicuously for at least ten miles, and we saw evidences of deepening of the first bottom deposits and slowing of current for fully thirty miles, above the dam. The cottonwoods of the first bottom within eight miles above the dam had all been killed, evidently by the raising of the general surface around their trunks; and the mesquites and other vegetation of the second bottom had all been drowned out, there thus being no trace of second-bottom conditions except for dead stalks. These were replaced by vast mud flats growing up to arrowweed. All this change, of course, involved the birds and mammals of the areas affected, in addition to the plant life. Below the dam reverse changes took place. The water, having dropped a considerable portion of its sediment above the dam because of the slowing of its current, was able to pick up sediment at a cor- respondingly accelerated rate below the dam. This, and the fact of a new cut-off having been found by the river in the delta in 1909, thus temporarily shortening its channel, resulted in a deepening of the channel seven feet below the previous level immediately below the dam. Thus the former flood-bottom was, in 1910, far above flood level, and in a way to become good second bottom, with appropriate metamorphosis in vegetation and fauna. Although these changes were local, and due to man’s interference, similar ones, due to natural causes, have doubtless occurred from time to time in various parts of its course in the river’s history, thus repeatedly shifting the riparian strips both in position and total width, with corresponding variability in the powers of the river at different 62 University of California Publications in Zoology | Vou. 12 points to act as a barrier to the dispersal of some animals and as a highway of dispersal for others. The effects of the extraordinary and continuous load of sediment of the Colorado River, together with the inconstaney of its channel, doubtless account directly or indirectly for many of the peculiarities in the fauna. As far as known to the writer, in the Needles-to-Yuma section of the river valley there are no aquatic molluses or decapod crustaceans, or tailed amphibians. Toads and frogs are present but not abundant. The fish fauna in the main stream is sparse in both species and individuals. Our party seined at three different points in the main stream. At two of these nothing was caught; in the third, a backwater slough on the Arizona side above Mellen, four sorts of fishes were taken, catfish (Ameiurus nebulosus), bony-tail (Gila elegans), hump-backed sucker (Yyrauchen cypho), and earp (Cyprinus carpio). A huge mimnow (Ptychocheilus lucius), ealled locally ‘* Colorado salmon,’’ was caught with hook and line in back-water on the California side opposite Cibola, and was plentiful immediately below the Laguna dam, where many were being taken by the Indians living near there. In lateral sloughs and overflow depressions carp and catfish were often observed in numbers, and in these relatively clear waters they were far more accessible to piscivorous animals than in the opaque water of the river itself. It may be said, in conelusion, that the portion of the Colorado River under consideration has relatively a very poor aquatic fauna and flora, and that this poverty has had its effect in limiting the occurrence of vertebrate animals dependent upon such sources of food-supply. ZONAL AND FAUNAL POSITION OF THE REGION Two schools of faunistie students are represented among American zoogeographie writers of the present day. One, of which C. H. Merriam (1894, ete.) is the most prominent exponent, sees in temperature the chief controlling eause of distribution, and deals with the ranges of species in terms of ‘‘life-zones.’’ The other school, of which C. C, Adams (1905, ete.), A. G@. Ruthven (1907), and Spencer Trotter (1912) are active advocates, assigns to temperature but a minor role, looking rather to a composite control, of many factors, resulting in ecologic ‘“associations,’’ of which plants are essential elements, and which are 1914] Grinnell: Mammals and Birds of the Colorado Valley 63 to be further explained on historical grounds. The two sets of areas thus defined do not by any means correspond. Yet the reviewer cannot fail to note, here and there, places where boundaries coincide, and such coincidences are so frequent as to be suggestive of real con- cordance in some significant manner. Is it not probable that both schools are approximately correct, the difference in mode of treatment being due to different weights given the different kinds of evidence, or, in other words, to difference in perspective? The opportunity is here taken to attempt to bring into aecord the systems of the two schools. The period of field study up to the present time devoted by the writer to the animal life of the climatically diversified state of Cali- fornia has led him to the recognition of three distinct orders of distri- butional behavior as regards terrestrial vertebrates. These are indi- cated in the terms: zonal, faunal, and associational. Every animal is believed to be limited in distribution zonally by ereater or less degree of temperature, more particularly by that of the reproductive season (see Merriam, 1894). When a number of animals (always in company with many plants similarly restricted) approximately agree in such limitation, they are said to occupy the same life-zone. The observation of this category of distributional delimitation is particularly easy in an area of great altitudinal diversity like that comprised in the southwestern United States. The writer is led to wonder if those authors who minimize the importance of temperature have ever been privileged to travel, and carry on field studies, outside of the relatively uniform eastern half of North America! Study of any area which varies widely in altitude and hence provides readily appreciable differences in daily temperature from place to place brings conviction of the very great effectiveness of temperature in delimiting the ranges of nearly all species of animals as well as of plants. Particular attention may be called to the results of a biological survey of Mount Shasta (Merriam, 1899). But temperature is not to be considered the only delimiting factor of environment, though its possible overemphasis by the Merriam school seems to have led some other persons to believe that this view is held. In fact, it becomes evident after a consideration of appro- priate data that very many species are kept within geographic bounds in certain directions only by an increasing or decreasing degree of atmospheri¢e humidity (see Grinnell and Swarth, 1913, p. 217). By 64 University of California Publications in Zoology — [Vou. 12 the plotting of the ranges of many animals as well as of plants, coincidence in this regard is found in so many cases as to warrant the recognition of a number of ‘‘faunal areas,’’ on the causative basis of relative uniformity in humidity. It is probable that every species is affected by both orders of geographic control. As to which is the more important, assembled data seem to show that more genera and higher groups are delimited by zonal boundaries than by faunal boundaries (see Merriam, 1892, p. 49, ete.) The arresting power of temperature barriers would therefore seem to be relatively the greater. In the third category of distributional control there is a con- spicuous association of the majority of so-called adaptive structures of animals (often of high taxonomic value) with certain mechanical or physical features of their environment. An animal may thus intimately depend upon certain peculiarities, inorganic or organic, or both, of a given area, and be unable to maintain existence beyond the limits of occurrence of those features of the environment. For instance, Dipodomys deserti is delimited by soil of certain texture and depth. Tracts of relatively uniform environmental condition, including their inanimate as well as living elements, are here called associations. The geographical distribution of any animal is correctly diagnosed in terms of each of the above three groupings. In other words, an animal belongs simultaneously to one or more zones, to one or more faunas, and to one or more associations. No one of these groupings can be stated in terms of the other, any more than a person can com- pute liquids by the peck, or weight in miles. The constituent species within each of these groupings always belong to the other two. To illustrate: the southern white-headed woodpecker inhabits the con- iferous forest association of the San Bernardino fauna of the Transi- tion zone; the Abert towhee belongs to the mesquite and quail-brush associations of the Colorado Desert fauna, of the Lower Sonoran zone; the Pacifie shrew belongs to the upland riparian association of the northern coast redwood fauna of the Transition and Boreal zones. Referring now to the region contiguous to the lower Colorado River, we have good reason, both biotic and meteorologic, for assigning it all to one zone, namely, the Lower Sonoran, and to one fauna, the Colorado Desert; but many associations are represented. In other words, the variation in altitude and latitude included is not great enough to bring sufficient modification of the characteristically high temperature to affect profoundly the distribution of the plant and 1914] Grinnell: Mammals and Birds of the Colorado Valley 65 animal life within the region. Nor is there marked variation in atmos- pherie humidity, the entire area being swept by air currents of pre- vailing dryness. To express the situation in a different way, zonal and faunal conditions are remarkably uniform; but associational econ- ditions are varied, as pointed out in the succeeding chapter. The zonal diagnosis, Lower Sonoran, is based in part upon the presence of the following determinative genera: MAMMALS BREEDING BIRDS Ammospermophilus Melopelia Sigmodon Micropallas Dipodomys Geocoecyx Perognathus Centurus Pipistrellus Calypte Macrotus Pyrocephalus Guiraca Phainopepla Toxostoma Auriparus Polioptila The faunal diagnosis, Colorado Desert, is based in part upon the presence of the following subspecies and species, selected from the entire list as being particularly restricted: MAMMALS BREEDING BIRDS Odocoileus hemionus eremicus Lophortyx gambeli Citellus tereticaudus tereticaudus Otus asio gilmani Peromyscus eremicus eremicus Bubo virginianus pallescens Reithrodontomys megalotis deserti Dryobates scalaris eactophilus Neotoma albigula venusta Centurus uropygialis Thomomys albatus Colaptes chrysoides mearnsi Dipodomys deserti deserti Agelaius phoeniceus sonoriensis Perognathus spinatus spinatus Melospiza melodia saltonis Perognathus penicillatus penicillatus Pipilo aberti Perognathus intermedius Piranga rubra cooperi Lepus ecalifornicus deserticola Vireo belli arizonae FRelis oregonensis browni Vermivora luciae Canis ochropus estor Dendroica aestiva sonorana Vulpes macrotis arsipus Toxostoma crissale Mephitis estor Polioptila plumbea Procyon pallidus Myotis occultus Myotis californicus pallidus The picked zonal and faunal ‘‘indicators’’ just named are often found in the same association with other elements seemingly less sensitive to temperature and humidity. Some of the latter, however, 66 University of California Publications in Zoology — | Vou. 12 may be present under duress, since their greatest abundance is known to be in adjacent zones or faunas. Thus Lynw and Eptesicus are much more plentiful in the Upper Sonoran zone or even in the Transition zone; and, on the other hand, Sigmodon and Pyrocephalus are genera of subtropical abundance. The presence of elements of the latter category was probably what led Merriam at one time (1894, p. 233, footnote, pl. 14) to refer the lower Colorado River Valley to the Tropical zone. All students who employ the life-zone system, now unhesitatingly agree in referring the area in question to the Lower Sonoran zone. The ‘‘western desert tract,’’? of Mearns (1897, pl. 2), is probably similar to the ‘‘Colorado Desert fauna’’ of the present paper. The former term is not considered apropos in the system adopted by the present writer on the ground that an inanimate area is designated thereby, rather than an assemblage of living things inhabiting the area. ASSOCIATIONAL AREAS OF THE REGION Since the entire region under consideration all belongs to one zone and all to one fauna, according to the definitions of these distributional terms given in the preceding chapter, the study of local distribution in the Colorado River region pertains chiefly to associations. Perhaps nowhere else in America can one find the degree of associational con- trast which is presented in the region under consideration. A stream of large volume, with paralleling strips of well-watered bottom land, maintains its course to the sea through what is considered the hottest and most arid desert in the world. There is nothing to show that the atmosphere is appreciably more humid in the vicinity of the bottom lands or the river itself than upon the open desert. The evaporated moisture is quickly dissipated ; that is, it becomes diluted to an imperceptible proportion in the desert air currents. It appears, therefore, that the great floral differences observed between the extreme associations are due primarily to differ- ence in amount of soil water available. There are, of course, such additional factors as varying alkalinity, and shade (see Spaulding, 1909). It must be understood that the associations here defined are recog- nized by the writer primarily beeause of their service in the treatment of animal distribution. Botanists have found it useful to make much 1914] Grinnell: Mammals and Birds of the Colorado Valley 67 finer analysis (for example, see Spaulding, 1908). A more or less detailed description of the associations observed in our 1910 explora- tions is herewith offered, as a necessary prelude to further discussion. In the lists of species, grouped according to apparent associational preferences, all mammals found by our party are included, and all the birds treated in the ‘‘General Accounts’’ except sixteen species, which are of transient occurrence and uncertain forage-ground like swallows, or else so rare as to make even approximate appraisement impossible. Qualifying terms are appended in each case: whether of exclusive (excl.) occurrence in the association under discussion, or of maximum (max.) abundance, or of minor (min.) abundance. Obviously a species of exclusive occurrence is entered in but one of the associational lists, while one marked mar. in one list is to be found marked min. in one or more of the other lists. It is to be taken for granted that in species of easy locomotion, individuals may occur in transit across other associations than the one or ones in which it is characteristically present en masse. There is necessarily more or less uncertainty in many cases, and where the doubt is strong, owing to lack of knowledge, a question mark is added. With birds, seasonal occurrence is indicated by the terms: resident (present throughout the year), winter (present throughout the winter, in certain cases including fall and spring also) , summer (present during the breeding season, in certain cases including spring and fall also), transient (merely passing through during migra- tion). River ASsocilaTION BIRDS Gavia immer: exel.; winter Erismatura jamaicensis: exel.; winter Sterna forsteri: execl.; transient Chen hyberboreus hyperboreus: exel.; Phalacrocorax auritus alboceiliatus: winter excl.; winter and transient Plegadis guarauna: excl.; transient Pelecanus erythrorhynchos: exel.; win- ter and transient Mergus serrator: exel.; winter Anas platyrhynchos: exel.; winter Nettion carolinense: exel.; winter Querquedula cyanoptera: exel.; tran- sient Spatula clypeata: exel.; winter Dafila acuta: exel.; winter Marila affinis: exel.; winter Mycteria americana: exel.; summer Ardea herodias treganzai: exel.; res- ident Butorides virescens anthonyi: exel.; transient Nycticorax nycticorax naevius: excl.; resident Grus canadensis: excl.; winter (?) and transient Fulica americana: min.; transient 68 University of California Publications in Zoology — (Vor. 12 Pisobia minutilla: exel.; winter Sayornis nigricans: exel.; winter Actitis macularius: exel.; winter Corvus corax sinuatus: max.; resident Oxyechus vociferus vociferus: min.; Petrochelidon lunifrons lunifrons: transient excl.; summer Cireus hudsonius: min.; winter Stelgidopteryx serripennis: min.; Pandion haliaétus carolinensis: exel.; summer transient Anthus rubeseens: exel.; winter Ceryle aleyon: excl.; transient MAMMALS Castor canadensis frondator: exel. Proeyon pallidus: max. Ondatra zibethica pallida: max. Remarks upon the River Association—For reasons already explained there is relatively little cryptogamic aquatic flora in the Colorado River. There is therefore little or no food-supply from this source to attract plant-eating ducks. This category of water-birds was, in facet, very sparsely represented. On the other hand, herons were notably plentiful because of the supply of catfish and carp made abundant at intervals by the drying-up of overflow ponds. While fishes were not abundant in the main stream, they were plentiful in backwater sloughs, where, too, the water was more nearly clear because the sediment had a chance to settle out. The ornis of the river appeared to owe its proportionate consistency in large measure to the above two circumstances, namely, poverty in aquatie plant life, and sporadie abundance of certain fishes in the lateral sloughs (see p. 62). The single carnivorous mammal (Procyon) belonging chiefly to the river association was piscivorous in food habits, foraging along mud bars (see pl. 4, fig. 2) and at the margins of overflow ponds and sloughs, as do the herons. The two rodents of the river and larger paralleling sloughs, which are here included as part of the River Association, lived in banks immediately adjacent to the water. The beaver fed chiefly upon bark, twigs and foliage of such willows and cottonwoods as had fallen over into the water through being under- mined by the current. 1914] Grinnell: Mammals and Birds of the Colorado Valley 69 WILLOW-Corronwoop ASssocIATION BIRDS Melopelia asiatica trudeaui: excl.; summer Accipiter velox: max.; winter Aceipiter cooperi: excl.; resident Buteo borealis calurus: min.; resident Falco sparverius phalaena: min.; resi- dent Otus asio gilmani: max.; resident Bubo virginianus palleseens: min.; resident Dryobates sealaris cactophilus: max.; resident Sphyrapicus varius nuchalis: min.; winter Centurus uropygialis: min.; resident Colaptes cafer collaris: exel.; winter Colaptes chrysoides mearnsi: min.; resident Phalaenoptilus nuttalli nuttalli: max.; winter (as a forager only) Phalaenoptilus nuttalli nitidus: max.; resident (?) (as a forager only) Chordeiles acutipennis texensis: min.; summer (as a forager only) Archilochus alexandri: max.; summer Tyrannus verticalis: excl.; transient Myiarchus cinerascens cinerascens: min.; transient Nuttallornis borealis: exel.; tran- sient Myiochanes richardsoni richardsoni: max.; transient Empidonax difficilis difficilis: exel.; transient Empidonax trailli trailli: exel.; sum- mer Empidonax hammondi: exel.; tran- sient Empidonax wrighti: excl.; transient Empidonax griseus: max.; winter Pyrocephalus rubinus mexicanus: min.; resident Molothrus ater obseurus: max.; sum- mer Xanthocephalus xanthocephalus: min.; winter Agelaius phoeniceus sonoriensis: max. ; resident Ieterus cucullatus nelsoni: max.; summer Teterus bullocki: max.; summer Zonotrichia leucophrys gambeli: min.; winter Spizella passerina arizonae: min.; transient (?) Melospiza melodia fallax: min.; win- ter Melospiza melodia saltonis: min.; resident Melospiza lineolni lincolni: min.; winter Pipilo aberti: min.; resident Oreospiza chlorura: max.; transient Zamelodia melanocephala melanoce- phala: max.; transient Guiraeca caerulea lazula: min.; sum- mer Passerina amoena: excl.; transient Piranga ludoviciana: max.; transient Piranga rubra cooperi: excl.; summer Vireosylva gilva swainsoni: max.; transient Lanivireo solitarius cassini: max.; transient Vireo belli arizonae: max.; summer Vermivora ruficapilla gutturalis: excl.; transient Vermivora celata celata: max.; win- ter Vermivora celata lutescens: excl.; transient Dendroica aestiva sonorana: exel.; summer Dendroica aestiva brewsteri: exel.; transient Dendroica aestiva rubiginosa: exel.; transient Dendroica auduboni auduboni: max.; winter Dendroica nigrescens: max.; transient Dendroica townsendi: max.; tran- sient Dendroica occidentalis: exel.; tran- sient Geothlypis trichas secirpicola: min.; resident 70 University of California Publications in Zoology [ Vou. 12 Teteria virens longicauda: exel.; sum- Polioptila caerulea obscura: min.; mer winter Wilsonia pusilla pileolata: max.; Hylocichla ustulata ustulata: exel.; transient transient Wilsonia pusilla ehryseola: min.; Hylocichla guttata guttata: excl.; transient winter Troglodytes aedon parkmani: excel.; Planesticus migratorius propinquus: winter max.; winter Regulus calendula cineraceus: exel.; winter MAMMALS Odocoileus hemionus eremicus: min. Uroeyon cinereoargenteus scotti: max. (?) (foraged in all other riparian as- Peromyscus maniculatus sonoriensis: sociations) max. Mephitis estor: min. Sigmodon hispidus eremicus: max. Procyon pallidus: min. Reithrodontomys megalotis deserti: Antrozous pallidus pallidus: max. min. (?) Sylvilagus auduboni arizonae: min. Myotis occultus: max. (?) Felis oregonensis browni: max. Nyectinomus mexicanus: max. (?) Remarks wpon the Willow-Cottonwood Association.—The predom- inating plants in this association were: willows of at least two species (Salix nigra and Salix fluviatilis), the cottonwood (Populus fremontr), and guatemote (Baccharis glutinosa). The latter occurred chiefly as an undergrowth where the willows or cottonwoods had reached large size. Practically all of the area occupied by this association is subject to inundation annually in early summer, of from a few inches to as much as twelve feet. Only such plants as can survive this period of drowning are able to oceupy the flood-bottom of the river. Furthermore, as explained in the general description of the river (p. 59), much of the overflow bottom is subject to destructive erosion through continual changing of the river’s channel. By this process all vegetation in its path is swept away at frequent intervals. Only such trees as are of rapid growth are able to maintain a representation on the major part of the bottom lands. No plants of the willow-cottonwood association (or it may be more briefly referred to as simply the willow association) occurred also on the upland deserts, and conversely no true desert plant occurred in the overflow area. Only one possible exception was observed, the serewbean (Prosopis pubescens) which occurred in mixture with wil- lows in a few places, where the flood-bottom was old, that is, had not 1914] Grinnell: Mammals and Birds of the Colorado Valley 71 been swept by channel-swinging for a long period of years. However, in the vicinity of the Colorado River, we failed to observe the screw- bean outside of the willow association, although it is known to occur in the mesquite and even the catclaw associations elsewhere; so that here the screwbean may be considered a restrictedly riparian plant. One other plant of the willow association requires mention—the cane (Phragmites communis), which grows in dense jungles on per- manent portions of the river bank from the vicinity of Picacho down. The tracts of cane are usually narrow, but, as in Canebrake Canon, below Picacho, may extend continuously close along the river for miles. At the higher stages of the river the bases of the stalks are submerged, while the drooping ends trail in the water (see pl. 4, fig. 3). A verdant screen on either hand thus intercepts the view of the voyager. The exogenous vegetation of the willow association is all of it deciduous, so that before the time of leafing-out in early March (see pl. 5, fig. 4) a very different appearance is presented from that later in the season. The spring growth is luxuriant, and accompanying it is a crop of insects which offer prolific food-supply to the abundant bird population of that season. Practically all of the birds listed for the willow association are either insectivorous or raptorial. Gramin- ivorous or spermophilous species are notably absent. The greater part of the fifty passerine birds listed for this asso- ciation are transients or winter visitants. Only three are permanent residents. The most notable characteristic summer visitants are: Dendroica aestiva sonorana, Vireo belli arizonae, Piranga rubra coop- eri, Archilochus alexandri, and Molothrus ater obscurus. Since the willow association includes practically the only forest elements in the region, we find here a concentration of dendrophilous birds, other than foliage-feeders, such as woodpeckers and flycatchers. The latter, in particular, are abundant in both species and individuals, and contribute to the great contrast afforded between the life of the riverside and that on the desert which in large part lacks them. While bird-life is conspicuously more abundant in the willow asso- ciation than in any one of the desert associations, just the reverse obtains with the mammals. With the exception of the insect-feeding bats, which share with the smaller birds the benefits of the insect supply, there is but one rodent of wide and plentiful occurrence— Peromyscus maniculatus sonoriensis. Three other rodents occur locally, notably Sigmodon. Otherwise the only mammals of the willow asso- ciation are far-ranging predators. =I bo University of California Publications in Zoology — [Vou. 12 The paucity of terrestrial mammals in this association is probably due to the repressive effect of the annual overflow which cannot fail to reduce the food-supply for many days at a time, even if extensive mortality does not directly ensue through drowning of individuals. As already implied (p. 58), the willow association varies greatly in width in different parts of the river’s course. Where the channel is constricted by rock walls, as in the box canon at The Needles, all trace of it is effaced for many rods. Where the river flows among hills patches of willows in ravine-mouths give detached representations to one or more elements. On the other hand, the broad valleys are occu- pied chiefly by this association which may then be as much as seven miles wide and continuous for many miles on one side or the other of the meandering channel. Taken by and large, the willow associa- tion is the most important one biotically of the entire set of associations dependent upon the presence of the river. TuLEe ASSOCIATION BIRDS Fulica americana: max.; winter Melospiza melodia saltonis: min.; Oxyechus vociferus vociferus: max.; resident winter Geothlypis trichas scirpicola: max.; Cireus hudsonius: max.; winter resident Xanthocephalus xanthocephalus: max. ; Geothlypis trichas oecidentalis: min.; winter transient Agelaius phoeniceus sonoriensis: Telmatodytes palustris plesius: excl.; min.; resident winter MAMMALS Sigmodon hispidus eremicus: min. Ondatra zibethica pallida: min. Reithrodontomys megalotis deserti: Procyon pallidus: min. max. Remarks upon the Tule Association—The river’s habit of overflow would be expected to result in rather extensive tracts of palustrine flora. As a matter of fact, however, marshes were few and of small size. This was probably due to the rapid rate of evaporation of overflow water so that favoring conditions did not last long, and also to the rapid silting-in of such water basins as ox-bow cut-offs. As a result there were either almost lifeless alkali depressions, or lagoons practically identical in biotie features with the main river. But in a few places there were well-defined palustrine tracts kept wet 1914] Grinnell: Mammals and Birds of the Colorado Valley 73 throughout the year, chiefly by seepage. These were always located back from the river near the outer edges of the broader valleys, where they were least affected during flood time. They were marked by erowths of tules, sedge, and salt-grass, sometimes the latter alone, and were usually surrounded by the arrowweed or willow association (see fig. B). The little open water sometimes attracted a few transient ducks and mudhens, but so far as known no water birds outside of the Ardeidae remain to breed anywhere along the Colorado River. As may be noted from the list, but few animals were found to frequent the tule association regularly enough to be considered dis- tinetive features of that association. Of these, Reithrodontomys was the only mammal finding its center of abundance there. ARROWWEED ASSOCIATION BIRDS Lophortyx gambeli: min.; resident Melospiza melodia saltonis: max.; (only as a forager) resident Geocoeeyx californianus: min.; resi- Melospiza lincolni lincolni: max.; dent (only as a forager) winter Astragalinus psaltria hesperophilus: Pipilo maculatus curtatus: max.; min.; resident winter Zonotrichia leucophrys leucophrys: Oreospiza chlorura: min.; transient min.; winter Vireo belli arizonae: min.; summer Zonotrichia leucophrys gambeli: Vermivora celata celata: min.; win- min.; winter ter Melospiza melodia fallax: max.; win- Dendroica townsendi: min.; transient ter Oporornis tolmiei: min.; transient MAMMALS Peromyscus maniculatus sonoriensis: Sylvilagus auduboni arizonae: min. min. Mephitis estor: min. Neotoma albigula venusta: min. Spilogale arizonae arizonae: min. (?) Perognathus penicillatus penicillatus: min. Remarks wpon the Arrowweed Association.—In its purity this asso- ciation possesses only one conspicuous plant, the arrowweed (Pluchea sericea), which, over extensive areas, grows so densely as to oceupy the ground to the exclusion of everything else (pl. 5, fig. 5). This belt of arrowweed usually occupies slightly higher ground than does 74 University of California Publications in Zoology — [Vou. 12 the willow association, yet not above the high-water mark. It is pres- ent quite regularly as a tract along the outer margin of the willow association, often only a few feet in width, but sometimes, as for example near Pilot Knob, as much as two hundred yards in width. The arrowweed grows to a very uniform height, varying in different places from three to over eight feet; and the slender straight stems stand so close together as effectually to bar rapid progress through the more luxuriant tracts. The plant is of perennial growth, but only during the season of blossoming, April and May, does it appear to be particularly favored by insects. In many places there is mixed willow and arrowweed, in fewer places quail-brush and arrowweed, either combination being evidently more attractive to birds than the arrow- weed alone. The only resident bird finding its maximum abundance in the arrow- weed association is the song sparrow (Melospiza melodia saltonis). Even here the suspicion is aroused that this association is sought more for cover because of the peculiar manner of growth of the constituent plant, than as a food-producing area. For the song sparrow forages freely into both adjacent associations. Of mammals, not one can be said to find its maximum abundance in the arrowweed association. The few species regularly trapped there were all more prominent constituents of adjacent associations. There would be little reason for the recognition of the arrowweed association as distinct from the willow association if it were not for its large extent and the conspicuous absence from it of a great many of the elements abundant in the latter. QUAIL-BRUSH ASSOCIATION BIRDS Lophortyx gambeli: max.; resident Zonotrichia leucophrys leucophrys: min.; transient Zonotrichia leueophrys gambeli: max.; winter Melospiza melodia fallax: min.; win- ter Melospiza melodia saltonis: min.; res- ident Pipilo maculatus eurtatus: min.; win- ter Pipilo aberti: min.; resident Oreospiza chlorura: min.; transient Guiraea caerulea lazula: max.; sum- mer Oporornis tolmiei: max.; transient Geothlypis trichas occidentalis: max.; transient Toxostoma crissale: min.; resident Polioptila plumbea: min.; resident =~ 1914] Grinnell: Mammals and Birds of the Colorado Valley MAMMALS Peromyscus maniculatus sonoriensis: Neotoma albigula venusta: min. min. Sylvilagus auduboni arizonae: max. Remarks upon the Quail-brush Association —The local hunters at Needles and other towns along the river apply the term quail-brush to Atriplex lentiformis, a plant growing in dense clumps three to eight feet in height along the outer margin of the flood-bottom. The belt of this plant is so well marked and continuous, and the important relations borne to certain animals are so evident, that the writer is led to designate it separately by the name quail-brush association (see pl. 6, fig. 6). Atriplex lentiformis often forms pure growths of a few yards to many rods in width on ground at about the upper limit of the average annual overflow. It was seen by us only in immediate prox- imity to the flood-bottom, and hence was essentially a riparian element, different from the other species of Atriplex inhabiting the region. The peculiar feature of the quail-brush was its compactly inter- lacing network of branches, so dense and resistent that a person might throw his whole weight against a thicket only to be hurled back by the elastic rebound. The value of the plant to the animals affect- ing it consisted pre-eminently in the protecting cover afforded. Quail and cottontail rabbits when frightened took refuge in runways or natural spaces beneath its tangled mass; and bush-inhabiting sparrows of several species seemed to find ideal shelter in it. Although serving thus as a temporary shelter to animals regularly foraging in other associations, the quail-brush association also pro- vided a safe breeding place for such birds as Pipilo aberti, Guiraca and Melospiza. In places, clumps of quail-brush alternated with mesquites, and here 7’orostoma crissale found particularly congenial ground. Mesquite ASSOCIATION BIRDS Lophortyx gambeli: min.; resident Centurus uropygialis: min.; resident Accipiter velox: min.; winter (only as a forager) Dryobates sealaris cactophilus: min.; Myiarchus cinerascens cinerascens: resident (only as a forager ?) min.; transient Sphyrapicus varius nuchalis: max.; Myiochanes richardsoni richardsoni: winter min.; transient 76 University of California Publications in Zoology Empidonax griseus: min.; winter Pyrocephalus rubinus mexicanus: max.; resident Molothrus ater obseurus: min.; sum- mer Astragalinus psaltria hesperophilus: min.; resident Astragalinus lawrencei: excl.; winter Zonotrichia leucophrys leucophrys: min.; transient Zonotrichia leucophrys gambeli: min.; winter Spizella passerina arizonae: min.; winter Pipilo aberti: max.; resident Piranga ludoviciana: min.; transient Phainopepla nitens: max.; resident Vireosylva gilva swainsoni: min.; transient Vireo belli arizonae: min.; summer Vermivora luciae: exel.; summer [ Vou. 12 Dendroica auduboni auduboni: min.; winter Dendroica nigrescens: min.; transient Wilsonia pusilla pileolata: min.; transient Wilsonia pusilla chryseola: min.; transient Oreoscoptes montanus: max.; winter Mimus polyglottos leucopterus: max.; winter Toxostoma crissale: max.; resident Heleodytes brunneicapillus eouesi: min.; resident Auriparus flaviceps flaviceps: min.; resident Polioptila caerulea obscura: min.; winter Polioptila plumbea: min.; resident Planesticus migratorius propinquus: min.; winter Sialia mexicana oecidentalis: exel.; winter MAMMALS Odocoileus hemionus eremicus: min. Peromyscus maniculatus sonoriensis: min. Neotoma albigula venusta: max. Perognathus penicillatus penicillatus: min. Sylvilagus auduboni arizonae: min. Felis oregonensis browni: min. (?) Lynx eremicus eremicus: max. (?) Mephitis estor: max. Myotis californicus pallidus: min. Remarks upon the Mesquite Association.—This association is given the vernacular name of what is perhaps the most widely known plant of the region, the mesquite (Prosopis juliflora). Along the Colorado River the mesquite is closely restricted to a rather narrow belt along the outer edge of the riparian area, mostly above the reach of the highest flood water. This belt is of course lacking where hills closely abut upon the river, save at the mouths of ravines. Along the great valleys it forms a nearly continuous tract consisting of straggling clumps or of well-formed, though small, trees, in the latter case some- times orchard-like in regularity of spacing. At no point did we see mesquites with trunks over eight inches in diameter. While evidently requiring abundant water at root, submergence of the ground for any length of time kills mesquites, as proven by 1914] Grinnell: Mammals and Birds of the Colorado Valley 77 the conditions above the Laguna Dam. Here the whole association had been effaced by drowning. On the other hand, the mesquite is unable to exist on the desert proper, even in the larger washes. Only at the mouths of these did scattering examples extend away from the actual river bottom, and then, in the most favored places, for not more than a quarter of a mile. Man’s occupancy of the region has affected the mesquite associa- tion more than any other. The great value of the mesquite trunks for fuel has led to its practical disappearance as a tree along much of the lower course of the river. The steamboats which once plied regularly between Needles and Yuma are said to be chiefly respon- sible for this depletion. Several pumping plants contributed to the demand for fuel. Mesquite trees are very slow of growth; tracts of stumps now mark many areas where luxuriant groves once stood. A considerable number of low-growing plants find places as elements in the mesquite association, but by far the most important to the animal life is the mesquite itself. This shrub, or tree, provides both shelter and food, the latter through its fruit and foliage (see pl. 6, fig. 7) either directly or by way of insects. The mesquite serves also as the host of a parasitic plant (pl. 6, fig. 6), a species of mistletoe (Phoradendron californicum), which when in blossom is visited by myriads of insects, and which produces an abundant and almost con- tinuous crop of berries. Several of the winter and resident birds of the mesquite association depend almost wholly on these mistletoe berries for their food. Notable among these are: Phainopepla, Mimus, Oreo- scoptes, Planesticus and Sialia. Four species of breeding birds (Pipilo aberti, Toxostoma crissale, Vermivora luciae, and Phainopepla nitens) find in the mesquite asso- ciation the center of their abundance. One mammal is characteristic of the same association, namely Veotoma albigula venusta. SALTBUSH ASSOCIATION BIRDS Lophortyx gambeli: min.; resident Phalaenoptilus nuttalli nitidus: min.; Zenaidura macroura marginella: resident (?) max.; winter Chordeiles acutipennis texensis: min.; Geococeyx californianus: min.; resi- summer (only as a forager) dent Sturnella neglecta: exel.; winter Phalaenoptilus nuttalli nuttalli: Poecetes gramineus confinis: excl.; min.; winter winter 78 University of California Publications in Zoology Passerculus sandwichensis nevadensis: excl.; winter Passereculus sandwichensis alaudinus: excl.; winter Zonotrichia leueophrys gambeli: min.; winter Spizella breweri: max.; winter [ Vou. 12 Amphispiza bilineata deserticola: min.; summer Amphispiza nevadensis nevadensis: max.; winter Lanius ludovicianus excubitorides: min.; resident (?) Thryomanes bewicki eremophilus: min.; winter MAMMALS Citellus tereticaudus tereticaudus: max. Peromyscus eremicus eremicus: max: Thomomys albatus: excel. Dipodomys deserti deserti: max. Dipodomys merriami merriami: min. Lepus ealifornicus deserticola: min. Sylvilagus auduboni arizonae: min. Taxidea taxus berlandieri: excl. (?) Myotis ecalifornicus pallidus: Max. (?) Pipistrellus hesperus hesperus: min. Perognathus penicillatus penicillatus: min. Remarks upon the Saltbush Association—Ranchers in the Colo- rado River region distinguish two portions in each of the valleys, the ‘‘first bottom’’ and ‘‘second bottom.’’ These are of course dupli- eated in reverse position on opposite sides of the river. The first bottom is the overflow area, and comprises all the associations treated up to this point in the present chapter, from the river to the mesquite association, inclusive. These associations together constitute the riparian belt proper. The second bottom is in the nature of a terrace or bench, and is situated above the reach of high water, extending from the mesquite association desertwards to the base of the mesa bluff. This second bottom is usually quite level and varies from a mere strip, of few yards in width, to a tract as much as a mile wide. It is for the most part clothed sparsely with xerophytic or halophytie vegetation, a predominating species in which is the saltbush (Atriplex polycarpa) ; hence the name adopted for the association represented. The soil is almost always of fine sand, often wind-blown. The pre- vailing westerly winds have caused a growth of sand-dunes on the southeast edges of the second bottom at several points, notably on the Arizona side above Mellen. The driving sand is often arrested about a bush and as the resulting dune grows, so does the bush com- mensurately. A serubby form of mesquite may thus constitute the core of a sand-dune. Several small plants are peculiar to these wind- formed hillocks, one of which is the sand verbena (Abronia villosa). 1914] Grinnell: Mammals and Birds of the Colorado Valley 79 Favorable conditions for burrowing here attract heteromyid rodents, notably the large Dipodomys deserti. A characteristic assemblage results which might be appropriately called the aeolian sand associa- tion. Its peculiarities are not, however, in the writer’s mind, sufficient to warrant giving it more than minor recognition. Future finer analysis may possibly justify the separate recognition of the aeolian sand association, especially when the reptiles of the desert at large are taken into consideration. Elsewhere on the second bottom, depressions frequently occur where rainwater from adjacent desert slopes leaves by evaporation more or less alkali. In cases of excessive deposition, bare white stretches result, without any vegetation at all. On somewhat less alkaline eround there may be tracts of Spirostachys occidentalis and Suaeda suffrutescens, both being shrubby plants popularly included under the names pickle-weed and iceplant. But the greater portion of the second bottom, as also a strip lead- ing back along either side of the desert washes, -is marked by the saltbush. This Atriplex is quite uniform in appearance wherever it grows, forming small but stout-branched bushes seldom more than two feet tall. The creosote bush (Larrea divaricata), although predominating in the next-deseribed association, occurs not infrequently as a minor element in the saltbush association. Sometimes individual plants of Larrea reach a very large size (pl. 7, fig. 8), much larger even than on the desert mesa. Where small alluvial fans from the higher adjacent mesa make down on to the second bottom, there is an influx of such plants as the coyote melon (Cucurbita palmata), rattle-weed (Eriog- onum inflatum), unicorn plant, locally called devil’s-claw (Martynia proboscidea), and sandbur (Franseria dumosa), all of which oecur on sandy parts of the upper mesa as well. Along shallow washes through the second bottom there are often extensive thickets of Lycium andersoni. In spite of the above indicated variations in floral constitution, there is remarkable homogeneity in the animal life of the saltbush association. In winter it is the preferred forage ground for a number of xerophilous fringillids, as shown in the accompanying lst of birds. The generally loose, sandy soil seems to be an attractive feature for burrowing rodents of limited fossorial powers such as Dipodomys mer- riami merriami. These also find abundant food in the residual seeds of numerous small annual plants which for brief periods thrive on 80 University of California Publications in Zoology the open ground between the shrubs. [ Vou. 12 One such plant, gathered exten- sively by Dipodomys deserti, is Achyronychia cooperi. Tt will be noted that the food-relations of the birds and mammals of the saltbush association and of the willow association are quite the reverse of one another; in the latter, insectivorous species prevail, in the former graminivorous or spermophilous. CREOSOTE ASSOCIATION (Masa) BIRDS Chordeiles acutipennis texensis: max.; summer Sayornis sayus sayus: min.; winter Spizella breweri: min.; winter Amphispiza bilineata deserticola: max.; Summer Amphispiza nevadensis nevadensis: min.; winter Piranga ludoviciana: min.; transient Salpinetes obsoletus obsoletus: min.; winter MAMMALS Ammospermophilus harrisi harrisi: min. (stony) Ammospermophilus leucurus leucurus: min. (stony) Citellus tereticaudus tereticaudus: min. (sandy) Peromyseus eremicus eremicus: min. (sandy) Thomomys chrysonotus: excel. Dipodomys deserti deserti: min. (sandy) Dipodomys merriami merriami: max. Perognathus bombycinus: exel. (sandy) Perognathus formosus: min. (stony) Perognathus penicillatus penicillatus: min. (sandy) Perognathus intermedius: min. (stony) Perognathus spinatus spinatus: min. (stony) Lepus californicus deserticola: max. Canis ochropus estor: max. (foraged at night practically everywhere else) Vulpes macrotis arsipus: excel. Myotis velifer: max. (?) Eptesieus fuscus: max. (?) Maerotus ealifornicus: max. (?) Remarks upon the Creosote Association (Mesa).—The creosote bush (Larrea divaricata) was found to be the most widely distributed shrubby species of all the desert plants (see pl. 10, fig. 14). It occurred in varying abundance from the second bottom and wash- sides to the tops of the highest hills. Only the most rocky hill slopes, and the periodically eroded wash-bottoms, lacked this plant altogether. Yet there were obviously preferred areas of growth, or, still more notable in this connection, areas where the creosote bush grew to the entire exclusion of all other ligneous vegetation. These areas, where 1914] Grinnell: Mammals and Birds of the Colorado Valley 81 Larrea was at least the most conspicuous plant, occurred mostly on the more level upland desert mesa. The accompanying distinctive assemblage of mammals justifies the recognition of what may therefore be appropriately called a creosote (or mesa) association. Tongues of typical creosote association often run up from alluvial slopes into the rough hill country, following ravines and terraces. On the other hand, as already noted, the creosote bush and some of its companions in places invade the saltbush association. Along desert washes the two associations may be blended in all particulars to such an extent that definite diagnosis as one or the other is difficult. As will have been observed from the lists, there are only two breeding birds of the creosote association proper, Chordeiles acutipennis texensis and Amphispiza bilineata deserticola; and neither of these are abundantly represented. But graminivorous mammals are plenti- ful in species as well as individuals. Not all the species, however, are found in exactly the same places. As elsewhere shown, identical ground on opposite sides of the river may possess distinct species because of the action of the river as a barrier. Furthermore, segregation of rodent population on the basis of ability to find or dig safe retreats is evident. The nature of the ground thus bears a controlling relationship. Parts of the desert mesa are swept clean of fine sand by the prevalent winds, the resulting sur- face consisting of packed gravel, or wind-worn pebbles (pl. 12, fig. 18). Other parts of the desert have a sandy soil; and in places accumula- tions of sand transported by the wind have grown into sand-dunes, having much the character of those described for the saltbush associa- tion on a previous page. On sandy ground a common small shrub was the sandbur (Fran- seria dumosa) ; the rattle-weed (Eriogonum inflatum) was ever sum- moning startled attention; and in the vicinity of Pilot Knob a species of Ephedra was common. On stony ground often no other plant was to be seen than the creosote bush; but everywhere remains of short- lived sporadic vegetation gave clue to the source of supply of the seeds upon which depended the rodent population. In the vicinity of Pot- holes and Pilot Knob the mesa association included scattering clumps of round-stemmed cactus, and ocotilla (Fouquieria). Each of the two physical types of ground here noted, namely, sandy and stony, possesses certain peculiar species of mammals, as well as of plants. This segregation is indicated in the list, and might again serve as basis for recognizing two separate associations, the 82 University of California Publications in Zoology (Vou. 12 sandy creosote, and the stony, or rocky, creosote; but by giving weight to similarities, as well as to differences, such distinction could be of but minor rank. Expediency seems to argue against further refine- ment in the present stage of study. Then, too, no such precise statis- tical gathering of data as regards the animals has yet been done as would warrant the exercise of such fine analysis. CarcLaw (OR WASH) ASSOCIATION BIRDS Lophortyx gambeli: min.; resident Zenaidura macroura marginella; min. ; winter Bubo virginianus pallescens: max.; resident Geococeyx californianus: max.; resi- dent Dryobates sealaris cactophilus: min.; resident Phalaenoptilus nuttalli nuttalli: min.; winter Phalaenoptilus nuttalli nitidus: min.; resident Archilochus alexandri: min.; summer Calypte costae: min.; summer Myiarehus cinerascens cinerascens: min.; Summer Ieterus cucullatus nelsoni: min.; tran- sient Ieterus bullocki: min.; transient Carpodacus mexicanus frontalis: max.; resident Astragalinus psaltria hesperophilus: max.; resident Zonotrichia leuecophrys leucophrys: max.; transient Zonotrichia leucophrys gambeli: min.; winter Zamelodia melanocephala melanoce- phala: min.; transient Phainopepla nitens: min.; resident Lanius ludovicianus excubitorides: max.; resident (?) Lanivireo solitarius cassini: min.; transient Oreoscoptes montanus: min.; winter Mimus polyglottos leueopterus: min.; winter Heleodytes brunneicapillus couesi: max.; resident Thryomanes bewicki eremophilus: max.; winter Auriparus flaviceps flaviceps: max.; resident Polioptila caerulea obscura: max.; winter Polioptila plumbea: max.; resident MAMMALS Odocoileus hemionus eremicus: max. Peromyscus eremicus eremicus: min. Dipodomys merriami merriami: min. Perognathus formosus: min. Perognathus penicillatus penicillatus: max. Perognathus spinatus spinatus: min. Lepus ealifornicus deserticola: min. Lynx eremicus eremicus: min. Uroeyon cinereoargenteus scotti: min. Pipistrellus hesperus hesperus: min. Remarks upon the Catclaw (or Wash) Association —Except for the Bill Williams and Gila rivers the Needles-to-Yuma section of the Colorado River receives no tributary stream, save as an immediate 1914] Grinnell: Mammals and Birds of the Colorado Valley 83 result of the very infrequent local cloud-bursts or severe thunder- storms. Even the two ‘‘rivers’’ named often go completely dry in their lower courses following protracted drouth. At frequent inter- vals along the valley of the Colorado, well-defined but ordinarily dry water-courses leading down from the adjacent uplands record the existence at some time or other of considerable run-off. Among the hills, deep, steep-sided ravines show the profound effects of erosion, even though, in this era, rains come but rarely. Where the water-course drains a large extent of higher country, enough to furnish a volume of water sufficient to find its way through the mesa to the river, it may occupy a broad valley with low confining banks or bluffs. Such a broad, dry flood plain is ealled locally a ‘‘wash’’; and as the biota of such ‘‘washes’’ is quite distinetive the term ‘‘ wash association’’ has been suggested (see pl. 7, fig. 9). Since a prevalent plant in its flora is the tenaciously thorny catelaw (Acacia greggi), the term catclaw association may be used as an alternative denomination. Both are appropriate terms. The vegetation of the catclaw association is the most conspicuous of all of the desert associations, for it includes several species which reach the stature of trees. The largest of these is the desert ironwood (Olneya tesota) which grows abundantly in all the larger washes on both sides of the river, from the lower Chemehuevis Valley at least to the vicinity of Picacho (pl. 8, fig. 10). The branches are leafy but thorny, forming ideal refuges for certain small birds and locations for their nests. The apparently leafless palo verde (Parkinsonia torreyana) is a close companion of the ironwood, and occurs also along the smaller ravines into the hills (see pl. 8, fig. 11; pl. 9, fig. 12). Both the catclaw and the palo verde were found in practically every wash on both sides of the river, from the vicinity of Needles to Pilot Knob. The smoke-bush (Dalea spinosa) is a conspicuous element in many of the washes from near Riverside Mountain to Pilot Knob. As an indication of the size reached by individuals of these truly xerophilous trees the following measurements taken by the writer, are here presented: An ironwood growing in a wash about one-half mile back from the river bottom in lower Chemehuevis Valley, Cali- fornia side of the river, was 90 inches in circumference of trunk two feet above the ground, and had a height of 31 feet. A palo verde nearby was 48 inches in circumference of trunk two feet above the ground, and was 28 feet in height. Another palo verde (pl. 9, fig. 12) growing in a wash on Californian territory four miles north of Pot- 84 University of California Publications in Zoology | Vou. 12 holes, was 60 inches in circumference of trunk two feet above the ground, first branch four feet above the ground, and total height 24 feet. It is thus apparent that birds of arboreal habit find only in the wash association of the desert a near approach to conditions preferred elsewhere, and this doubtless accounts for the conspicuous transient arboreal element occurring in this association. But the greatly reduced foliage, giving most of the above-named trees the aspect of winter leaflessness, results in close resemblance to brush or shrubbery, as indicated by the prevalence of the brush-inhabiting category of breed- ing birds affecting them. The ironwood oceasionally harbors clumps of mistletoe (Phoraden- dron californicum) in common with the mesquite of the riparian belt. This accounts for the presence of certain berry-eaters. There are also berry-producing shrubs bordering the washes, notably Lyciwm ander- soni and Lycium parishi. The former occurs widely as a character- istic member of the catelaw association. The latter, a much larger thicket-forming shrub, was noted only in small washes in the vicinity of Picacho. As already noted, Atriplex polycarpa occurs in an exten- sion of the saltbush association leading up along each side of nearly all of the larger washes. Bird-life is better represented in the catclaw association than in any other of the desert associations. Some of the species are closely adherent to it, being evidently by structure and habits dependent upon the conditions pertaining to thorny brush. But mammals are relatively less numerous. Only one rodent finds its maximum abund- ance along the washes (Perognathus penicillatus penicillatus, see fig. D), and it is possible that even of this species, the metropolis is in the adjacent sandy saltbush tract, and that trapping really waylaid the individuals foraging at large away from their homing places. Times of deluge, even if of rare occurrence, are doubtless accompanied by great mortality of ground-dwelling mammals along these washes. This factor must be one of no small import in determining the biotic constitution on the several levels of the desert surface. 1914] Grinnell: Mammals and Birds of the Colorado Valley 85 Saguaro ASSOCIATION BIRDS Faleo sparverius phalaena: max.; Colaptes chrysoides mearnsi: max.; resident resident Otus asio gilmani: min.; resident Myiarchus cinerascens cinerascens: Micropallas whitneyi: exel.; resident max.; Summer Centurus uropygialis: max.; resident Carpodacus mexicanus frontalis: min.; resident MAMMALS (As far as known, same as in creosote association.) Remarks upon the Saguaro Association —The conspicuous columns of the giant cactus or saguaro (Cereus giganteus) first met our expect- ant gaze just below the mouth of Bill Williams River. There, on both sides of the Colorado, for a stretch of two miles or more, stood many specimens growing on the hill slopes in full view from the river as we floated by. A landing was made on the California shore, and sev- eral saguaros closely examined for nesting sites of birds. We next found these cactuses on the open desert one to three miles east of the river on Arizona territory around Ehrenberg. There were so many here that they formed a distinct tract, extending across the mesas and occupying the interlying washes as well. Giant cactuses again came to view some ten miles below Picacho, two or three individuals being seen on the California side and many on the Arizona side. Finally, on the mesa on both sides of the river, three to five miles above the Laguna dam, saguaros were plentiful. On the California side from one hundred yards to two miles back from the outer edge of the river bottom were about seventy-five individuals (see pl. 9, figs. 12, 13). Thirty-seven, big and little, were counted by the writer as in sight from one point. It was gratifying thus to find this unique plant well represented on California ground. For coming with it into the state of California were the several birds listed, an assemblage altogether justifying the designation of a saguaro association. This association is best devel- oped in a large area of southwestern Arizona. The tongue crossing the Colorado River above Laguna and Potholes is undoubtedly an exten- sion of it. As far as we were able to learn, in other plant elements and in mammals, the saguaro association was here identical with the creosote 86 University of California Publications in Zoology [ Vou. 12 association. It is not improbable, however, that further work would disclose the presence of a number of species of both mammals and birds, at least on the Arizona side, not now known so far west. It is clearly apparent that the critical feature of the saguaro which prescribes its avian dependents is nothing else than the favorable opportunity offered for the excavation of safe retreats in its trunk. And only the two woodpeckers are equipped for making these exea- vations. So that, without the woodpeckers to make holes, the other birds would be no better off for the presence of the saguaro. As it happens, at least one of the species of woodpeckers (Centurus uropy- gialis) invariably accompanies the cactus. This is a most interesting form of contingent or incidental inter- dependence of animal and plant. Since the giant cactus during the ereater part of the year produces no fruit and harbors no insect life, it follows that all the birds using its cavities as roosting or nesting places glean their livelihood from the surrounding desert. ‘lhe latter, as already stated, presents conditions which seem to be practically identical with the creosote association, which is so nearly barren of bird life, and of vast extent beyond the limits of the saguaro. One form of associational restriction is hereby proven, namely, that by dependence upon safe home retreats. Encexia (Rocky Hints) Association BIRDS Buteo borealis calurus: max.; resident Faleo mexicanus: exel.; resident Bubo virginianus pallescens: min.; resident Aéronautes melanoleucus: excl.; resi- dent Calypte costae: max.; summer Sayornis sayus sayus: max.; resident Corvus corax sinuatus: min.; resident Carpodacus mexicanus frontalis: min.; resident Astragalinus psaltria hesperophilus: min.; resident (?) (as a forager only) Spizella passerina arizonae: max.; winter Spizella breweri: min.; winter Spizella atrogularis: exel.; transient Amphispiza bilineata deserticola: min.; summer Stelgidopteryx serripennis: max.; summer Salpinetes obsoletus obsoletus: max.; resident Catherpes mexicanus conspersus: excl.; resident (?) MAMMALS Ovis canadensis nelsoni: exel. Ammospermophilus harrisi harrisi: max. Ammospermophilus leucurus leucurus: max. Peromyscus ecrinitus stephensi: excel. 1914] Grinnell: Mammals and Birds of the Colorado Valley 87 Neotoma intermedia desertorum: Corynorhinus macrotis pallescens: excl. excl. (?) Perognathus formosus: max. Pipistrellus hesperus hesperus: max. Perognathus intermedius: max. (foraging everywhere else) Perognathus spinatus spinatus: max. None of the Remarks upon the Encelia (Rocky Hills) Association. hills or ‘‘mountains’’ in the near vicinity of the lower Colorado River is of such great altitude as to bring a reduction in temperature to an extent sufficient to modify its biotic complexion. The tallest of The Needles rises to less than 2,000 feet above the level of the river. In other words, as far as observed, the plants and animals of the hills show no distributional behavior other than as explained on associational grounds. But the hill country does exhibit a distinctive association of plants and animals, setting them apart sharply from the desert mesa, or the riverside. A conspicuous shrub on the rocky steeps, especially at The Needles, was the dense Encelia farinosa, with its light gray foliage, growing on talus slopes and even in crevices of the cliffs (pl. 10, fig. 15). Other plants of the same locality were: Atriplex confertifolia, Larrea divaricata (sparsely intermixed), Asclepias subulata (visited regularly by the Costa hummingbird), Fagonia californica (on the hottest slopes of broken rock), Hyptis emoryi (a ‘‘sage-bush’’ five to six feet high growing on the sides of ravines and at time of blossoming, in March, frequented by hummingbirds), Perityle emoryt (an abund- ant composite annual, the seeds of which were much sought after in March by fringillids), and Miihlenbergia debilis (a grass growing in shaded ravines and providing forage for graminivorous rodents, like Perognathus). While to distant view the hills seemed more barren of vegetation than most of the other associational areas, nearby inspection showed abundant remains of inconspicuous annual plants. This in part would seem to account for the great numbers of mammals present, as shown by our trapping. The additional favorable factor was evidently the abundance and availability of natural retreats afforded in the talus and fractured outcrops. With insectivorous and raptorial species, such as the swifts and bats, hawks and owls, the rocky hills served merely as home retreats, foraging being doubtless carried to the more productive lowlands, 88 University of California Publications in Zoology — (Vou. 12 } Nedoma deserlorum Enceha Perognalhus inflermedius creosole bush \Ammospermophilus herrisi wNOoZ/YY Dpockbrys merriami salt bush Perognalthus penicillatus Peromyscus erermicus mesquile Sy/vilagus arizonae quail brush Peromyscus sonoriensis errowweed Neolorna venusta willow sapere eg, tte Kafe 4aAlby opesyjo? Jo yfo4y jeuojras /eapy wIA4 OOVHOTOS Vv Sy suoyelrossy jeuiuley Pe /Ue/q BuUIMOYS willow arrowweed Peromyscus sonoriensis Neolorma venusta gual brush mesquile Peromyscus erermnicus Sylvilagus erizonze salf bush Perognalhus penicillalus Dipedomys merriam creosole bush de y bINYOS/TYD Ammospermophilus leucurus Perogralhus spinalus Perosnalhus formosus Encelia Neolorma deSerforum | Peromyscus slephensi 1914] a ‘Su suoyerrossy /euiuey PYe yu Grinnell: Mammals and Birds of the Colorado Valley &d BUIMOYS 2 Sea Sa//ius Ly, n ~ = => z A BS OYyrery fo ¢ lef yes? fo aife4y4 jeuoyres /°% LON ves /J@4 49A HINOIS 2OWIIIS = FA OOVHOTOD NOISSIAAIT 170M IV MNS IO/S VINHOHTVD fiber ‘Caslor willow Reiltrodonlomys Ile Fiber willow Neolma venusla proaned Peromyscus arenas le brush | mesquile Cesc eremmicu. agus arizonae saltbush dhol ANC mesquile Perognahus pencilolus ) saltbush ’ Dpodomys merriami | | Ammospermophilus leucurus creosole bush | Wectova deste fore | acorns spinatus rognhus formasus 89 90 University of California Publications in Zoology — (Vou. 12 particularly along the river bottom. Only two insectivorous birds were restricted to the Encelia association, namely, the wrens, Salpinctes obsoletus obsoletus and Catherpes mexicanus conspersus. Both are by structure adapted to gleaning from crevices of rock surfaces after the manner of nuthatches on tree trunks. GENERAL DISCUSSION OF ASSOCIATIONAL RESTRICTION From the preceding description of the conditions in the region studied, it is obvious that there are two groupings into which all the designated associations can be classed, namely, riparian and desert. The riparian set of associations includes those which owe their pres- ence to the existence of the river, and is delimited outwardly at the mesquite association (see figs. A and B). The Colorado River appar- ently exerts no influence beyond the immediate bottom lands, which are affected by the underground water supply. All of the desert set of associations are represented in varying proportion over the vast arid tracts stretching away to the east and west of the river. Often they are discontinuous, but recur again and again in the same fauna with the same constitution. The riparian associations are thus narrow strips of varying width closely paralleling the river from north to south and persisting prac- tically continuously from the point of emergence of the Colorado River from the Grand Canon to the Gulf of California. In the broad delta region the riparian associations spread out so that there are great areas of each, doubtless sufficient to be computed by the square mile. It is thus possible to trace the elements severally, of each association, to places of prevalence over considerable areas, even though those ele- ments are, on the upper river, scattered sparsely along a narrow strip. Associational diagnosis of species thus often becomes possible when a knowledge of local conditions alone would be inconclusive. This prin- ciple deserves enlarging upon. There were caught in the same trap-line opposite The Needles both Peromyscus eremicus eremicus and Peromyscus crinitus stephensi. It might have been impossible to say from the data gathered at that particular point, where the adjacent associations were complexly inter- mixed, just what sort of ground each species preferred. But trapping previously done in the salt-bush association in the Imperial Valley 1914] Grinnell: Mammals and Birds of the Colorado Valley 91 showed the presence there of P. e. eremicus only, while field-work among the rocky hills in the vicinity of Victorville, on the Mohave Desert, showed P. c. stephensi only. As another illustration, along the Colorado River, Melospiza melodia saltonis and Pipilo aberti were often found on common ground, although evidently averaging differently in associational preference. To prove beyond doubt what is the true ecologic niche of each, a knowledge of the distribution of each species elsewhere in their respec- tive ranges becomes necessary. In the extensive arrowweed tracts around the west end of Salton Sea, Welospiza melodia saltonis is an abundant species while the other bird is absent. In the mesquite belt not far distant to the west, in the vicinity of Martinez and Torres, Pipilo aberti is prevalent, and the song sparrow absent. Hence the species of towhee in question may be confidently assigned to the mes- quite association, and the subspecies of song sparrow to the arrowweed association. It is not to be inferred that all species behave in this clearcut fashion associationally, any more than that all do so zonally or faunally (see Grinnell and Swarth, 1913, p. 220). At the same time the writer feels fairly sure of adequate grounds for proposing a general law in this regard, namely, that where the faunist happens to meet with a heterogeneous assemblage of biotic elements, not subject to clear asso- ciational diagnosis in the restricted locality of first observation, assign- ment of the species each to a well-defined association becomes possible by tracing out their ranges severally into the adjacent areas. A coneurrent axiom is that if associational analysis is carried far enough, no two species of birds or mammals will be found to oceupy precisely the same ecologic niche, though they may apparently do so where their respective associations are represented fragmentarily and in intermixture. In determining the associational status of mammals we have to deal chiefly with elusive animals, of nocturnal habits, which are hidden away during the day for the most part beyond reach. Trapping is not an altogether certain index to association; for individuals may be caught repeatedly in a trap-line which may not happen to intersect at all the regular forage ground or breeding home of the species. Indivy- iduals forage far and wide beyond the limits of their home territory and at the close of the breeding season wander in similar fashion. Some species, including nearly all mammals except xerophilous rodents, regularly travel far for water. 92 University of California Publications in Zoology [ Vou. 12 Attention is here called to our records of the capture of pocket mice. An unexpectedly large number of species of the genus Perognathus was found to occur in the region traversed. As many as three species were taken in one night in traps placed close together. But by testing many localities and comparing the results we soon came to know where to expect each separate species. The diagrams herewith presented (figs. C-F) show in statistical form the associational preferences of four species of Perognathus. Providing the same number of trap- nights (counting one trap set one night as one “‘trap-night’’) was Ais SN Are p ¢ = Tl «g Ce | = Se oa Si) ax > Ee) Ee Ere MS TOU pc Hl Ue ES wet ES RE say ie oe OM eat aS oO 4” eR Gos pe Ren Se RB Ss Bh en is oe eo = = eee 3 H fe de et eo ee Ores Fig. C. Diagram showing frequency of capture of Perognathus formosus in the several associations. There were fifty-three individuals trapped, of which the associational occurrence was satisfactorily recorded. devoted to each association, this method should be fairly accurate. There will here oceur to the reader ways of securing much greater precision in results of field-work in the future. But in fact, as our experience grew, far more effort was expended on ground promising additional specimens of rare species, than on ground already thor- oughly exploited. While a mammal or bird may be closely confined to a narrow territory characterized by certain conspicuous plants, the critical factor or factors of its environment may be quite apart from food requirements. As an illustration, the case of Dipodomys deserti 1914] Grinnell: Mammals and Birds of the Colorado Valley 93 deserti may be cited. This rodent, taking its entire, irregular and discontinuous range into consideration, is closely restricted to those portions of the desert affording a deep surface layer of fine sand. Areas of aeolian sands constitute its typical home territory (see pl. 11, fig. 16). Now, it is quite probably not any peculiarity of food-supply g a 3 S ce c=) = S a al eo 2 J es od 2 =] 2 a 1S) o 4 eS = & ao a = 2 Pa) = 5 2 EI s = 6 fs 6 Se Se Bae iI s Dn <= 4 ov i im A = ® = 3S 2 a i =a = si a te a fe i > YHATY OGVYOTOD YAMOT AHL AO AAIS VINYOAITVO Gout 4O[1d qouyT 4OUd qOusy 40[td ‘UN “Tu G ‘AN. “Yur g ‘OU'N “Tur ¢ ‘aN ‘Tur ¢ “SUN ‘rut ¢ "aN “Tu G apl9A O[eg apl9A O[eg eplaA Oleg AqyRBoory pep ‘ant P ‘pep ped ‘pup "pe p “ped ‘pe Pp ped pep ‘pe Pp ‘pep ady xag STLOT 98901 Seo0L FE9OL 6e90L 6690L TE90L O0€90T 6690T 83901 LG90L 9690T ‘ou ‘snyy HHL WOU SQ0INdTIA SAdIdSIH NOGOWDIS JO SYALUWITTIIN NI SINAWANOSVAW GNV LSI 1914] Grinnell: Mammals and Birds of the Colorado Valley Doo On the other hand, these aberrant examples might be recognized as mutants, that is, variants of phylogenetic significance though not geographically isolated. After going over the ground, I lean per- sonally towards the notion of the existence of an exceptionally wide range in ordinary fluctuational variation, as accounting for the extremes in question. This is coupled with the inadequacy of the material available, so that the distinctions are emphasized. With large series more nearly normal frequency would be expected. In this connection it may be pointed out that an alleged race, Sigmodon hispidus arizonae Mearns (see Bailey, 1902, p. 108), has been described from Fort Verde, Yavapai County, Arizona, with char- acters close to those appertaining to the Colorado River giants in question. In size these two specimens from the Colorado equal or exceed that of the type of arizonae, so that the existence of a separate race based on size is questionable. Reithrodontomys megalotis deserti J. A. Allen Desert Harvest Mouse This rodent was confined strictly to the riparian strips. Although not found by us above Ehrenberg, it probably does occur in suitable environs as far up the Colorado as these are afforded, that is, up to the steep-walled cafons above Fort Mohave. Specimens were trapped close to the water’s edge, usually in grassy places adjacent to tracts of willows. In some cases these mice were on islands cut off by chan- nels from connection with the shore. In practically all cases the ground they occupied would at high water be entirely submerged, so that the mice would either have to swim to higher ground or take refuge in the willows and drift piles. Doubtless they swim freely and are also swept back and forth across the river channels at flood time. Twenty-one specimens were preserved (nos. 10165-10185). The localities of capture were: Arizona side: Ehrenberg, five miles north of Laguna, and Yuma; California side: near Palo Verde, and five miles northeast of Yuma. Young over half grown were taken May 2, and on the same day a female containing five embryos. Neotoma albigula venusta True Colorado Valley Wood Rat This species of wood-rat was in distribution the exact associational complement of Neotoma intermedia desertorum: it was restricted to 234 University of California Publications in Zoology [Vou. 12 the riparian strips, and no evidence was forthcoming of its oceurrence out on the desert mesa, in the hills, or even up the desert washes. As compared with rodents of the desert proper this species might for the sake of emphasis be termed semi-aquatic, for it was trapped within a few feet of the main river, and its tracks were noted on fresh mud at the water’s edge. Individuals were caught on willow-grown islands, which were subject to submergence with rising water, so that the animals would be forced to live in trees or take to water. No signs of nests were observed in trees anywhere, and it seems reasonable to suppose that these wood rats voluntarily swim narrow channels, especially when pressed by hunger. Although foraging widely over the bottomlands, the permanent abodes of these rats appeared to be chiefly located at the lateral rims of the riparian belts, just at the upper reach of high water. This, too, marked the belt of mesquites, so that the mesquite association can be confidently assigned as the distinctive habitat of Neotoma a. venusta. At no point did we find the ‘‘enormous nests’? of this species deseribed by Mearns (1907, p. 474) as found by him below Yuma. His name, Neotoma cumulator, for this species was selected on this account. A few small stick houses were found by us in the mesquite strip above Mellen, and nests of similar scant proportions, at other places down the river. In all cases the rats appeared to have ready access to subterranean burrows. At the mouth of the Gila River, near Yuma, many burrows were found in a tract of guatemote without trace of stick nests, and a number of the animals were caught. It is quite possible that recurring unusually high flooding discourages the rats in the portion of their range explored by us from amassing much material, since it is subject to being floated off. We found small young as early as February 24 and 25 at Mellen, and from that date on to May 5, near Yuma. It is probable that the breeding season is just about closed by the time the annual overflow begins, so that young as well as adults are able to seek safety for the brief period of exile by assuming arboreal habits or swimming to higher ground. The series of 68 specimens of Neotoma albigula venusta obtained (nos. 10463-10530) represent localities of capture as follows: Cali- fornia side: near Riverside Mountain, 2; Palo Verde, 2; twenty miles above Picacho, 8; eight miles below Picacho, 2; five miles northeast of Yuma, 8; Pilot Knob, 3. Arizona side: Mellen, 12; Ehrenberg, 3; twenty-five miles below Ehrenberg, 1; ten miles below Cibola, 7; five 1914] Grinnell: Mammals and Birds of the Colorado Valley 235 miles northeast Laguna, 6; near Yuma, 14. Two adult males from twenty miles above Picacho and near Yuma weighed nine ounces each; an adult female from the former station weighed seven ounces. No differences are observable between the specimens from the two sides of the river. The river is evidently in scarcely any degree a barrier to distribution in this species. Rather has the Colorado bottom served as a highway of invasion for the species from its center of distri- bution, which is manifestly to the southward, the desert on either side acting as confining walls. An examination of the series of specimens shows an apparent slight decrease in size up the river, that is, away from the assumed center of dispersal. But since there is also con- tinued, though diminished, growth of individuals with age, and because of the few examples from any one of the uppermost stations, this direction in variation up the river is not to be considered as established. Much more material is needed. Neotoma intermedia desertorum \erriam Desert Wood Rat The 41 specimens taken (nos. 10424-10462, 10716, 10717) indicate stations of occurrence as follows: California side: opposite The Needles, 3 specimens; Riverside Mountain, 2; opposite Cibola, 1; twenty miles above Picacho, 13; eight miles below Picacho, 3; Potholes, 2; Pilot Knob, 11. Arizona side: foot of The Needles, 3; above Bill Williams River, 1; ten miles below Cibola, 2. None of our specimens was trapped within the riparian strips; all were taken on the desert proper, though at The Needles, where because of the abrupt rock walls the riparian element is in places reduced to a mere nearly vertical band, desert wood-rats or signs of them were found within a few yards of the river on the opposite sides. Here scanty accumulations of sticks were observed in crevices among loose boulders on hillsides or in clefts of the walls of ravines. At Riverside Mountain, some two miles back from the bottomlands, nests of large size were noted among boulders; this was true, too, at Pilot Knob, where also sign was seen and the rats themselves caught around clumps of desert tea far out on the mesa. Fifteen of the specimens taken at the different stations are young of varying sizes. Two very small young of 172 and 167 mm. length, respectively, were trapped March 6 and 7 at The Needles, indicating early breeding. Others nearly as small were taken at Pilot Knob, May 236 University of California Publications in Zoology | Vou. 12 10; so that there is not such constancy in breeding time in this species as in certain other rodents of the region. A careful comparison of the six specimens secured from the Arizona side with the much larger series from the California side, shows both cranially and as regards external features, no tangible differences which might be expected to occur (and do oceur in other cases) in a mammal of the desert proper and whose range is divided by the Colorado River. It appears that this species of wood-rat has not previously been found to the east of the Colorado River (see Goldman, 1910, pp. 76, 77). In fact, in his revision of the genus Neotoma, Goldman, in dis- cussion of Neotoma lepida and its subspecies stephensi (which are forms not dissimilar to desertorum), states (1910, p. 80) that ‘‘the ranges of the two [lepida and desertorum] are completely separated by the effective barrier of the Colorado River.’’ While the results of our work detract from this statement as to fact, the implication remains the same, namely, that the Colorado River may have been at one time of prime service in effecting the isolation of the lepida stock, particu- larly in the upper (northeastern) course of the river (see Goldman’s map, p. 77). I sent my Arizona-side examples of desertorwm to Mr. Goldman with the request that he examine them closely and give his opinion as to their possible approach in characters to N. l. stephensi. The latter form is, by the way, the nearest wood-rat of the same group recorded from east of the river, its nearest station being (Goldman, 1910, p. 80) the Hualpai Mountains, Arizona, only about fifty miles east of The Needles. Mr. Goldman replied that my specimens were ““typical desertorum,’’ and hence not bearing any significant resemb- lance to stephenst. It would appear, therefore, that the desertorum, as found by us on the Arizona side of the Colorado at the three stations named, is of probably direct and relatively recent descent from the stock on the California side. While this wood-rat does not inhabit the river bottom- lands, as does Neotoma a. venusta, it is quite possible that individuals forage down to the water’s edge where the riparian strip is reduced in width or practically wanting, as at The Needles. In event of rats becoming marooned on drift-rafts at periods of rising water such animals would be in a position to be ferried across the river; for it is not unlikely that such rafts would be carried by the swerving currents to opposite banks. No matter where such waifs should be landed they 1914] Grinnell: Mammals and Birds of the Colorado Valley 237 would, ultimately, barring accidents, tend to reach the same sort of environs they were used to, namely, the rocky hill slopes. While such a series of propitious events is in but a remote degree possible, it appears to me the most logical way of explaining the extension of the range of desertorum across the Colorado to the Arizona side. As above intimated, this occurrence, if a fact, does not militate against Goldman’s hypothesis that lepida and desertorum may have been held apart by the Colorado River, and are still so held apart in most of its course. Ondatra zibethica pallida (Mearns) Pallid Muskrat Evidence, hearsay and direct, indicates the presence of muskrats all along the Colorado River, from above Needles to below the Mexican line. They occur both in the main stream and in the various diverging sloughs of the big valleys. Our nine specimens (nos. 10652-10660) were all secured on the California side of the main stream, three near Palo Verde and six near Pilot Knob. No signs of houses were seen anywhere, the muskrats appearing to resort entirely to holes in banks where the current was sluggish. Near Palo Verde, April 1, a system of burrows was dug out by Stephens and Jones. This system of holes was in the bank of a tule-bordered slough about a half mile above its confluence with the main river. The entrances all opened considerably beneath the surface of the water at its level at the time. From these the burrows sloped upwards as they extended back from the slough, until those parts farthest from the water were two to three feet above its level at that stage. Some green tule stems were found in the passage-ways. One of the blind leads was filled in with packed tule stems, mostly dead ones, but moist and crushed. The regular nests consisted of dry tule stems, some of them shredded, and laid two to three inches deep. Some of the short blind leads showed fresh claw scratches in the earth at their ends. The passage-ways were ordinarily five inches in diameter, in places more, and usually kept near the surface of the ground, follow- ing its irregularities pretty closely. In places there was not more than three inches of earth above the burrow, and there was seldom as much as a foot. The system was in fact discovered by one of the party’s stepping upon a thin place and breaking through. The impression gained from a survey of the uncovered burrows was that the excavator had at the beginning burrowed up from the 238 University of California Publications in Zoology (Vou. 12 slough five to eight feet, then gone back a third of the way and started a branch to the water for another entrance. After finishing that, it had come half way back on the branch and dug another branch from it inland, then a branch from that to the water again. Roughly speaking, a series of connected Y’s thus resulted, with ends alter- nately in the water and in the bank landwards. It would appear as though such a plan of runways were well adapted to eluding enemies, both terrestrial and aquatic. An adult male and two two-thirds-grown young were trapped in the entrance burrows. In the vicinity of Pilot Knob tLe muskrats were living along the steep northern bank of the main river, where the water was overhung by a dense growth of cane. A number of willow branches sagging into the water and drift logs caught in the tangle of cane (pl. 4, fig. 3) showed themselves to be the nightly rendezvous of muskrats. On these, at the farthest projecting portions, and usually not over four inches above the surface of the water, capsule-shaped pellets of excre- ment to the number of three to a dozen or more marked the perching places of the rats. This excrement was usually fresh, as the logs would go awash with the frequent cross-river winds. Number 0 steel traps set on these logs without attempt at concealment, either with or with- out bait (apple or potato), caught five adult muskrats, May 10 to 14. Four out of the five were drowned, the traps being provided with chains which in turn were nailed to the logs. A half-grown youngster was shot on the latter date as it swam among some trailing cane stalks. Muskrats have recently invaded the Imperial Valley along the irrigation canals leading from the Colorado River. I visited this valley in February, 1912, and was told there that the California Devel- opment Company found it necessary to hire men to shoot and trap muskrats because of the damage done in burrowing through the levees. The rats even follow the smaller ditches out into the farms at Calexico, Heber, and El Centro. At Calexico, February 8, I obtained a fresh specimen shot by a boy on a farm close by, on the California side of the Mexican boundary. The entire series of ten specimens of the pallid muskrat in the Museum displays the now well-known features peculiar to the race represented, namely, small size, pale color and relatively scanty pelage (see Mearns, 1907, p. 495). The weights of the five full-grown male animals were: 31, 24, 21, 26, and 27 ounces, respectively; average 26 ounces. 1914] Grinnell: Mammals and Birds of the Colorado Valley 239 Thomomys chrysonotus Grinnell Ehrenberg Pocket Gopher The only place on the Arizona side of the river where gopher sign was seen was on the mesa within two miles back from Ehrenberg. Here two sets of mounds were found on low sandy ridges and a single gopher caught from a burrow, March 27. This specimen is a male, young adult (no. 10617) and proved so different from any previously described race of gopher as to warrant making it the basis of a new specific name (see Grinnell, 1912, p. 174). This is evidently an upland species, that is, not riparian, and its nearest relationships are suggested to be with forms to the east and north rather than with T. perpallidus or T. albatus of the desert region west of the river. Thomomys albatus Grinnell Imperial Valley Pocket Gopher At only one station on the California side of the river did we find any sign of gophers. This was at our last camp, east of Pilot Knob, where on the site of the old Hanlon Ranch several sets of work- ings were located, and eight specimens taken (nos. 10618-10625), May 7 to 10. Four of the animals were young, not more than half grown. These workings were all on the first bottom just at the outer mar- gin of the arrowweed association, in ground barely reached at the highest level of overflow. The absence of gophers in the greater por- tion of the Colorado bottom is reasonably explained by the occurrence of the yearly overflow which would drown them out. The colonies invading the river bottom do so only at points which they reach from some more favorable center of distribution back from the river. As already shown (Grinnell, 1912, p. 172), our gophers from the Pilot Knob station belong to the species occupying the alluvial delta region including Imperial Valley, which species is distinct from both the one represented on the Arizona side at Ehrenberg (Thomomys chrysonotus) and the one of the western end of the Colorado desert, at Palm Springs (Thomomys perpallidus ). The characters distinguishing albatus from chrysonotus cannot be reasonably considered as due alone to the action of the river as a barrier, for the former species belongs to a different association (saltbush) from the latter which, as far as known, adheres to the 240, University of California Publications in Zoology [Vou 12 creosote association. Thus the two species might exist on the same side of the river, with complementary ranges. Still, as far as known, neither species occurs on the opposite side of the river, the river thus serving as the factor of absolute delimitation in the respective cases. It is probable that the two species had latterly a separate course of origin, having been derived from distantly loeated stocks. Dipodomys deserti deserti Stephens Big Desert Kangaroo Rat The thirty-nine specimens secured were preserved as follows: nos. 10352-10381, skins with skulls; nos. 10708, 10709, complete skeletons ; nos. 10724-10730, aleoholies. The following localities are represented : Arizona side: Mellen, 23; Ehrenberg, 2; twenty-five miles below Ehren- berg, 1; five miles northeast of Laguna, 3. California side: Riverside Mountain, 1; near Pilot Knob, 9. This rat shows strong preferences for the aeolian sand association. It was found present practically wherever such accumulations reached an extent of an acre or more of sufficient depth to contain the burrows. This essential depth was seen to be at least a foot. Because of the limited powers of digging, as shown both by weak incisors and small front feet and claws, only such loose and fine-grained substratum could be occupied by this rodent. The packed floor of the mesa or the hard- pan of the hillsides was uninhabitable. The presence of this Dipodomys could always readily be recognized by means of conspicuous burrows, the mouths of which were usually at all times open. The looseness of the ground in which the burrows are dug results in their frequent caving in when men or horses walk over them. When the wind had not acted to efface them, the tracks of the rats could be plainly seen on the surrounding sand surface (pl. 11, figs. 16, 17). These tracks showed imprints of the hind feet and tails, indicating the tripedal mode of ambulation characterizing the Heteromyidae. The associational restriction of this large kangaroo rat gives the impression that it possesses a colonial habit. But it was observed that where proper conditions were continuous, as at the north base of Pilot Knob, the local distribution over the suitable area was practically uninterrupted. The burrows oceur in small groups, each group prob- ably representing but a single home center and occupied at times by a single adult, at others by an adult pair, and at others by adult and 1914) Grinnell: Mammals and Birds of the Colorado Valley 241 young. Tracks were to be seen far and wide between the scattering eroups of burrows. At Mellen, February 25, specimens of Dipodomys deserti trapped had their cheek-pouches filled with the minute blossoms and stem ends of Achyronychia cooperi, an inconspicuous plant growing prostrate upon the surface of the broad, sandy intervals between the creosote and Atriplex polycarpa bushes (pl. 11, fig. 16). The series of specimens from Mellen, on the Arizona side at the north, is faintly darker in color dorsally on an average than the series from near Pilot Knob on the California side at the south. Otherwise I can see no differences between specimens from the two sides of the river. Like Peromyscus eremicus and Perognathus penicillatus, Dipo- domys deserti inhabits the second bottom in places along the river, though much less continuously than these smaller rodents. Doubtless, as with them, individuals occasionally forage on to the first bottom, and thus run the chance of being transferred from one side of the river to the other by means suggested in other parts of this paper. Dipodomys merriami merriami Mearns Merriam Kangaroo Rat A series of 168 specimens of this small kangaroo rat were pre- served, and besides these many were discarded. The Museum numbers are: 10186-10351 (skins with skulls), 10710-10712 (skeletons), 10753— 10741 (alcoholics). The following localities are represented: Arizona side: Mellen, 18; foot of The Needles, 6; above Bill Williams River, 6; Ehrenberg, 10; twenty-five miles below Ehrenberg, 3; ten miles below Cibola, 8; five miles northeast of Laguna, 12; Yuma, 2. California side: five miles below Needles, 3; opposite The Needles, 4; Cheme- huevis Valley, 10; near Riverside Mountain, 6; above Blythe, 3; oppo- site Cibola, 19; twenty miles north of Picacho, 21; eight miles below Picacho, 12; Potholes, 5; Pilot Knob, 20. As indicated above, this was a widespread and abundant desert rodent, being taken in nearly every line of traps outside of the over- flow bottom. Its greatest abundance occurred on sandy ground adjacent to desert washes, on tracts of aeolian sands on the desert mesa, and on the second bottom paralleling the river on either side through the broad valleys, but wholly above the reach of the highest floods. In a number of cases specimens were trapped on the packed 242 University of California Publications in Zoology — [Vou. 12 ground of the upper mesa, and occasionally on rocky hillsides. These instances probably show great extent in foraging radius, the home burrows being in the softer ground of neighboring ravine bottoms or in depressions where sand could lodge. For, as far as observed, the burrows are always dug by the animals themselves, and their burrow- ing powers are weak. Their small size, however, does not require anywhere near the depth of workable soil that the huge D. deserti needs. Probably the far greater restriction in distribution of the latter species compared with that of D. merriami is determined by this difference in depth of soil required. The burrows of D. merriami are not easy to locate, as the entrances are left smoothly closed during the day. But where the fine wind-laid sand composed the surface of the ground, the hind foot and tail tracks showed clearly the routes taken by the rats to and from their burrows. The breeding season is at its height in April, though the following dates show considerable latitude. Females with embryos were taken on February 19, March 5 and 6, April 1 to 4, 19 and 21, and May 8. On each of the above dates but one animal is concerned, except during the period from April 1 to 4, when observations were made upon at least ten, and April 19, upon two. In sixteen cases there were two foetuses, in two cases three. The above data indicate a relatively slow rate of reproduction: but one litter per year is raised and there are seldom more than two young to a litter. This indicates a much safer existence, individually, for this kangaroo rat than for the ground squirrels and white-footed mice of the same habitat. In the case of these latter rodents litters consist of four to eight young, and in some species, at least, two or more litters are born each year. Kangaroo rats must be relatively very successful in escaping the numerous enemies that assail the rodent population of the desert. The writer finds it impossible to refer the Colorado River series of Dipodomys merriami, as a whole, to the subspecies D. m. simiolus (type locality, Palm Springs, on the western arm of the Colorado Desert), as might have been expected. The material shows great range in variation; some of the specimens, particularly in the series from the lower course of the river, are, it is true, very close to simiolus. On the other hand many, especially from the Arizona side of the river above Ehrenberg, are not with certainty distinguishable from near- topotypes of D. m. merriami, from south-central Arizona. It proves, 1914] Grinnell: Mammals and Birds of the Colorado Valley 243 however, impossible to assort the series into two subspecies on any satisfactory basis, and, since the average appears nearest D. m. mer- riami, all are placed under this name. Perognathus bombycinus Osgood Yuma Pocket Mouse Twenty-five specimens were procured (nos. 9956-9977, 10742— 10744), three being preserved as alcoholics, and the rest as skins- with-skulls. Only two localities are represented, near Ehrenberg on the Arizona side of the river, where eighteen specimens were caught March 25 to 30, and near Pilot Knob on the California side of the river, where seven were taken May 7 to 15. Those at Ehrenberg were caught on areas of wind-blown sand, especially where heaped about the bases of creosote bushes, on the mesa a mile or more back from the river. At Pilot Knob the species was found on aeolian sand accumula- tions and in shallow sandy washes on the desert mesa. In all cases Perognathus bombycinus was found on common ground with Dipodomys deserti, Dipodomys merriami, Peromyscus eremicus, and Perognathus penicillatus. The five rodents named thus have very similar associational preferences. While Perognathus bombycinus was not found by us as near the river as the second bottom (as were all the other rodents named) it is fair to assume that it may so occur where conditions favor. And like the others of the same association, transfer of individuals from side to side is likely to have taken place at inter- vals in the past. This might be advanced as a reason for the close similarity of the representations of bombycinus on the two sides of the river. The original description of Perognathus bombycinus (Osgood, 1907, pp. 19, 20) was based on a single specimen from Yuma, Arizona, and two from just over the Mexican line in Sonora. Our present series bears out to the dot the cranial characters assigned by Osgood, namely, as compared with Perognathus panamintinus bangsi, enormously inflated audital bullae and mastoids, the latter conspicuously pro- tuberant posteriorly, and narrow interparietal. Externally our speci- mens of bombycinus are distinguishable from bangs? in their very pale coloration, which consists in lighter ground color (dilute pinkish buff ) and almost obsolete black tippings to the hairs dorsally. Although, as compared with bangsi, bombycinus has greatly enlarged audital bullae, the external ear is not of appreciably greater size. 244 University of California Publications in Zoology [Vou. 12 The measurements of twenty adult males from our series are, aver- ages and extremes: total length 139 mm. (130-149), tail vertebrae 79 (70-85), hind foot 18.5 (17-19.5). I am unable to find any differences between the specimens from Ehrenberg and those from Pilot Knob, the species being thus identical on the two sides of the river. We found no traces of pocket mice of this group anywhere else along the river. Osgood (1907, p. 20), how- ever, records typical bangsi from Needles, California. It is quite pos- sible that there is a hiatus between the ranges of bangsi and bomby- cinus. At any rate, among the large series of bangsi in the Museum from the Mohave desert and from the west side of the Colorado desert from San Gorgonio Pass to the Mexican line, there is no individual showing close enough approach in characters to bombycinus to war- rant classing it as an intermediate. Although the affinities of bomby- cinus are clearly with the panamintinus group, there is good evidence for carrying the former as a full species. Curiously, only one out of the entire twenty-five bombycinus taken was afemale. This, and the fact that no young-of-the-year were found, would seem to show that up to May 15 the breeding season had not yet begun. For experience with other rodents in different places indicates that as soon as the young are born the females forage abroad actively, and are then caught in at least equal proportions to the males. Perognathus formosus Merriam Long-tailed Pocket Mouse A series of 44 specimens was obtained, 40 skins-with-skulls (nos. 9652-9691) and 4 alcoholics (nos. 10794-10797). These were taken exclusively on the California side of the river and represent localities as follows: five miles below Needles. 6; opposite The Needles, 12; Chemehuevis Valley, 8; Riverside Mountain, 5; near Blythe, 1; Pot- holes, 5; Pilot Knob, 7. The known range of this species is thus carried southeast along the west side of the Colorado River to the Mexican line (see Osgood, 1900, pp. 40, 41). The distribution is not, however. continuous, for marked restriction is shown to hilly country and rough mesas. The immediate valley of the Colorado is avoided, as also the broad. fiat and low desert depressions between the mountain ranges. Yet within the rough country trapping showed that Perognathus formosus occurred most frequently upon the narrow strips of loose, 1914] Grinnell: Mammals and Birds of the Colorado Valley 245 sandy soil along ravine bottoms or along washes cutting the elevated mesas. This closer associational phase of occurrence is significant on comparison with the local distribution of Perognathus spinatus. For- mosus and spinatus both live west of the river, and both occupy the same general associations (see figs. C, F). But locally the two are often found on separate ground, spinatus adhering most closely to the vicinity of boulder-strewn hillsides and fractured outcrops, while formosus prefers smoother ground, as noted above. At Riverside Mountain formosus was caught on the mesa, but not next to the abruptly rising base of the mountain where spinatus was found. But opposite Cibola, where pocket mice were phenomenally abundant, examples of Perognathus spinatus, P. formosus and P. peni- cillatus were all caught within a radius of fifty feet. The first two species were found in a single trap on successive mornings. As further considerations with regard to the occurrence of for- mosus and spinatus on common ground, it is to be remembered that the two species belong to separate subgenera (Perognathus and Chae- todipus, respectively), and thus have structures so different that impor- tant differences in food or other requirements are suggested; so that the forms are probably complementary instead of violently competitive. The breeding time of Perognathus formosus is indicated by the capture of blue-pelaged young at Potholes and Pilot Knob on April 28 and 29, and May 10 to 14. Preceding the first-named date no signs of breeding were in evidence, so that relative lateness in this annual function is indicated. Perognathus penicillatus penicillatus Woodhouse Colorado Desert Pocket Mouse A series of 196 specimens preserved, 184 as skins-with-skulls (nos. 9692-9875), two as skeletons (nos. 10720, 10721), and ten as alco- holies (nos. 10798-10807). The following localities are represented: California side: five miles below Needles, 8; opposite The Needles, 10; Chemehuevis Valley, 17; near Riverside Mountain, 16; near Blythe, 2; above Ehrenberg, 1; opposite Cibola, 35; twenty miles above Pic- acho, 10; eight miles below Picacho, 8; Potholes, 1; four miles below Potholes, 1; five miles northeast of Yuma, 5; near Pilot Knob, 8. Arizona side: Mellen, 12; above Bill Williams River, 2; Ehrenberg, 16; twenty-five miles below Ehrenberg, 3; ten miles below Cibola, 27; five miles northeast of Laguna, 8; near Yuma, 6. 246 University of California Publications in Zoology [| Vou. 12 In addition, the Museum contains four skins (nos. 5636, 5637, 5825, 5826) taken in 1861 by Dr. J. G. Cooper at Fort Mohave, on the Arizona side twelve miles above Needles. It is at once patent that of the five distinct species of Perognathus found along the lower Colorado, P. penicillatus is both most abundant and most widespread. Very many more individuals were trapped day by day than were preserved. It may be asked why explicit record is not here presented, not only of specimens saved, but of all the individuals trapped. The reason is that so similar in external appearance are the various species of Perognathus, that especially with our previous unfamiliarity with them, we were liable to many mistakes in identification in the field. After museum study was made of the suites of specimens brought home, with skulls cleaned and measurements tabulated, it was found that errors had actually occurred in a number of cases thus checked up. Perognathus penicillatus and P. intermedius had been confused on the Arizona side, and P. penicillatus and P. formosus on the California side, especially in the cases of immatures. Field determinations were therefore totally disregarded. It may searcely be necessary to remark that penicillatus (plus angustirostris, if the form designated by the latter name be consid- ered tenable), intermedius, spinatus, and of course formosus, are all distinguishable with absolute certainty (for characters see Osgood, 1900). No difficulty was experienced in allocating all specimens as soon as cleaned skulls were available. Although obtained at all collecting stations and from all associa- tions, reference to the graph for frequency of occurrence (see fig. D) clearly shows that Perognathus penicillatus has marked preferences for one particular environment. This preferred habitat is characterized by ground of fine-grained sand. The second bottom along the river and the broad washes of the flatter parts of the desert furnish the ideal conditions; and in the most typical of these penicillatus is either the exclusive or the prevailing member of the genus present. Penicillatus is the only species invading alkali depressions at the edges of the broad valleys back from the river (where Suaeda grows), and is the only species occurring regularly on parts of the overflow area of the river bottom. Referring again to the frequency record (fig. D), it is shown that penicillatus was found in every member of the riparian group of associations, even into the willows. Individuals were always, however, as far as record and memory serves to estab- 1914] Grinnell: Mammals and Birds of the Colorado Valley 247 lish the point, trapped on dry surfaces, indicating a preference for dryness. On thoroughly dried-out, rather high sand banks along the river where the bottom strip was narrow, sign was noted to within a few feet of the edge of the swift current. Individuals were trapped on sand bars, at the time high and dry, but so slightly elevated that rising water would at first form islands of them and then engulf them completely. No doubt, as with Peromyscus, aggressive individuals of Perog- nathus penicillatus are often thus caught and set adrift, thus giving a chance for transfer from one side of the stream to the other. As discussed in a general way elsewhere, there seems to be significant cor- relation between the fact of this free invasion of the river bottom and existence and similarity of the representatives of the species on both sides of the river. The breeding season of penicillatus is of greater duration than that of the other pocket mice of the region, as shown by the following data. A female taken April 19 contained five embryos; another taken May 4 contained four embryos. Gray-coated young-of-the-vear were taken as early as Mareh 5 and at intervals from then on through April and into May. After study of the Colorado River series in connection with the 120 additional specimens in the Museum from various localities in the Salton Sink region of the Colorado desert, I have come to the conclusion that the recognition of the name angustirostris as apply- ing to any of the Colorado River specimens is impracticable. Osgood (1900, pp. 45-47) deseribed a race angustirostris from the Colorado desert, type locality Carriso Creek, west side of the desert. The Museum has a good topotype series, and also a series from Mecea, in the bed of the Salton Sink at the northwest end of Salton Sea. The characters assigned to ‘‘angustirostris’’ were, as compared with penicillatus, small size, less massive skull, and longer and more slender rostrum. Osgood referred his few specimens from the Colorado River, from Ehrenberg and above, to penicillatus, while those from Fort Yuma and all localities to the west he listed under angustirostris. Close scrutiny of the extensive material now at hand shows great instability in all the alleged characters. There is as notable fluctuation in size from place to place up and down the river, as across the Colorado desert. The topotypes of angustirostris include some skulls as massive as many from Mellen, Arizona. 248 University of California Publications in Zoology — [Vou. 12 One thing is apparent, that the Mecca series includes more small specimens, with narrow-snouted skulls, than series of like extent from elsewhere. There is thus a tendency towards the existence of a race, of the characters assigned by Osgood to his angustirostris, in Salton Sink. But if the upper Colorado Valley specimens are penicillatus, so also are those from all the way down to Yuma and Pilot Knob and thence across to the east flank of the Coast Range in eastern San Diego and Riverside counties. In view of the locally fluctuating variations through the region under consideration, the inclusion of all under the one name penicil- latus seems now the wisest course. Should the penicillatus of Wood- house prove to represent a truly distinct race, occupying the elevated north central deserts of Arizona, as hinted by Osgood, then angustiros- tris would be the name to be used for the subspecies of penicillatus occurring all along the Colorado Valley below the Grand Cafion, and through the Colorado desert. Perognathus intermedius Merriam Intermediate Pocket Mouse A series of 82 specimens taken, nos. 9876-9955, saved as skins- with-skulls, and nos. 10785, 10786, as aleoholics. The localities of cap- ture were all on the Arizona side of the river, as follows: Mellen, 19; foot of The Needles, 10; above Bill Williams River, 2: Ehrenberg, 13; twenty-five miles below Ehrenberg, 1; ten miles below Cibola, 27; five miles northeast of Laguna, 10. The preferments of this pocket mouse are to all appearances iden- tical with those of the Harris ground squirrel. Both rodents are restricted to rocky hills and the hard-surfaced, coarse-graveled mesa. In both situations scattering creosote bushes form the prevailing vege- tation, though on the hills Encelia farinosa is an additional conspicuous plant. Perognathus intermedius was in no case found so near the river as the second bottom (see diagram, fig. E). In fact, only three out of the 82 individuals caught were found in sandy desert washes, and these could have readily reached the points of capture in foraging down from home centers on adjacent hills or mesa. Our series of record stations, as above named, carry the known range (see Osgood, 1900, pp. 52, 53) of P. intermedius to the west, and mark its limits in that direction as being at the east side of the 1914] Grinnell: Mammals and Birds of the Colorado Valley 249 Colorado River bottom. The presence of the species was ascertained at every station where appropriate ground was trapped, from Mellen nearly to the Laguna dam. Among all the specimens taken, only two were juvenals. These were taken above Laguna on April 24 and 25. None of the numerous females taken earlier in the season contained embryos, so that the breeding season would appear to be deferred in the case of this species until the advent of hot weather. Perognathus spinatus spinatus Merriam Spiny Pocket Mouse This pocket mouse was found only on the California side of the river. A series of 125 specimens was taken, nos. 9536-9651 (skins- with-skulls), 10713, 10714 (skeletons), 10787-10793 (alcoholies). Localities are represented as follows: opposite The Needles (practically topotypes of the species), 14; Chemehuevis Valley, 1; Riverside Moun- tain, 26; opposite Cibola, 36; twenty miles above Picacho, 32; eight miles east of Picacho, 6; Potholes, 4; Pilot Knob, 6. The spiny pocket mouse proved to have much the same associational preferments as the antelope ground squirrel, namely, the hills and rough-surfaced mesas of the desert. It was found close to the river only where the riparian bottomland associations were pinched in to merest traces of their elements by the abutment of the hills. Places of this nature, where spinatus was caught within as near as a hundred yards of the water’s edge, and yet on ground perfectly appropriate to the species, were opposite The Needles and twenty miles above Picacho. At the latter locality the only exception to the above statements eame to notice. An adult female was caught among the willows on first bottom and within seventy-five feet of the water, hence where proper conditions of topography and rising water might have resulted in transporting the animal to the opposite side of the river. But the excessive rarity of such a combination of circumstances probably accounts for the fact that Perognathus spinatus has never been found to the east of the Colorado River. The strip of bottomland at the point where this wanderer was captured was only about three hundred yards wide, and the species was found to occur commonly on the hillsides down to the outer edge. The individual probably occurred merely as a forager. 250 University of California Publications in Zoology (Vou. 12 Opposite Cibola, as already noted under P. formosus, the spiny pocket mouse occurred on a broad desert wash, strewn with boulders. Specimens were trapped there, not only among boulders but around scraggly stumps of ironwood. The breeding season apparently does not begin till April. On the 13th and 19th of that month females taken contained four embryos each. On April 28 and May 9 half-grown young were trapped. Lepus californicus deserticola Mearns Colorado Desert Jack Rabbit The eight specimens secured (nos. 10682-10689) came from the following localities: California side: five miles below Needles, twenty miles above Picacho, and near Pilot Knob; Arizona side: Ehrenberg, and five miles north of Laguna. Specimens from the two sides of the river are identical as far as I can see; and this is to be expected. For although the jack rabbit belongs primarily to the desert mesas, it forages also all over the flood-bottoms. I was told that this species of rabbit swims strongly when forced to take to the water. Jack rabbits occurred through the region in 1910 in very small numbers. At some stations none at all were seen. though old sign was usually plentiful. The greatest numbers were observed on the mesa back of Ehrenberg, but not more than half a dozen could be seen during a three hours’ circuit. Two females shot near Ehrenberg, March 28 and 29. contained two and three embryos respectively. The weight of an adult male (no. 10682) was four pounds, four- teen ounces; of an old female (no. 10683) six pounds. Sylvilagus auduboni arizonae (J. A. Allen) Arizona Cottontail The year of our exploration of the Colorado Valley was stated by the residents of the region to be an off-year for both jack rabbits and cottontails. Whatever the usual conditions. both kinds of rabbits were unexpectedly scarce. The cottontails were chiefly riparian in their local distribution. In fact. the only individuals seen desertwards beyond the mesquite belt were in the salt-bush association closely adjacent. and (opposite The Needles) up a catclaw wash not to exceed one-half mile from the river. 1914] Grinnell: Mammals and Birds of the Colorado Valley 251 The quail-brush association was the most preferred local habitat, the rabbits finding ideal refuges there beneath the thickets of Atriplex lentiformis. They ranged all over the bottomlands, and since individ- uals were seen on islands already cut off by rising water from main- land connection, it is probable that many rabbits are washed adrift each year and are carried from side to side of the river. We were assured that cottontails had been seen swimming in the river during flood time. We procured twenty-one specimens of the cottontail (nos. 10661- 10681), representing the following localities: Arizona side: Mellen, foot of The Needles, five miles northeast of Laguna; California side: opposite The Needles, Riverside Mountain, Blythe, opposite Cibola, eight miles east of Picacho, five miles northeast of Yuma, near Pilot Knob. S “ae a] ~— Ee ii eels ale = z & 2 Be 2 £8 Z an os 2s 3 a g as E SE 4 bh 2 x Ss % & $2 Z 2 Fs eS i ko pa Fa D rc) Q NS As Zz BAS bP a Ch Sees ee ee 43.5 48 144 21.7 3 198 161 127.5 36.0 41 19 134 22.5 1 Measured from plane of inferior surface of palatines vertically to highest point on surface of frontals. Lynx eremicus eremicus Mearns Desert Wildeat Two bobeats were trapped by our party: a female (no. 10604) in a patch of screwbean trees in the river bottom on the California side five miles below Needles, February 21; and a male (no. 10605) in a wash lined with palo verde and eatclaw on the Arizona side north of Mellen, February 27. In addition to these specimens, both saved as skins and complete skeletons, there is in the Museum a skin with the skull inside (no. 5620) taken by W. W. Holder at Mineral City (=Ehrenberg), Arizona, in 1864. or of dorsal 254 University of California Publications in Zoology [Vou. 12 The three skins agree in relative pale tone of coloration as com- pared with appropriately selected material from the Pacifie slope of southern California (—=Lynx eremicus californicus). The paleness consists In more extended white-tipping to the hairs. The Mellen specimen is slightly the darkest; but there is as much or more varia- tion in skins from a single locality elsewhere. The measurements, in-the-fiesh, of the two taken in 1910 are: Total Tail Hind Height Tuft of Mus. no. Sex length vertebrae foot of ear hair on ear 10604 2 815 150 175 63* 257 10605 rol 780 150 170 75 26 * Measured on dry skin. There are in the Museum three skins of wildeats from Victorville on the Mohave Desert. which are also somewhat pale in comparison with Pacifie Slope skins. Aside from this character of paleness, I am unable to find any other diagnostie features in wildeats of the desert region, even after carefully following Mearns’s deseription (1897, p. 457). Eremicus and californicus are certainly very close, though present material appears to me to warrant separate recognition (see Stephens, 1896, pp. 210, 211). The weight of the male wildcat caught near Mellen was 18 pounds. The stomach of this animal contained the fragments of at least two wood-rats. These had not been chewed finely, but had been chopped up into chunks about an inch long; the feet were still entire. Canis ochropus estor Merriam Desert Coyote Coyotes were only fairly common along the Colorado Valley. They were occasionally seen on the desert mesas by daylight, but were rarely heard. It was evident from tracks on the mud bars that they foraged down to the river’s edge at night. The six specimens secured (nos. 10611-10616) represent the fol- lowing localities, all on the California side: five miles south of Needles, opposite The Needles, twenty miles north of Picacho, and Pilot Knob. The weights of three adult males were 16, 18, and 21 pounds; of two females, 18 and 20 pounds. 1914] Grinnell: Mammals and Birds of the Colorado Valley 255 Vulpes macrotis arsipus Elliot Desert Kit Fox This fox appeared to be a characteristie element in the purely desert fauna. The five specimens taken were all trapped on the desert mesas back from the river. In fact, no evidence at all was forthcoming to show that kit foxes ever visit the river or even the bottomlands. Their presence was generally detected in sandy tracts, especially around colonies of Dipodmys deserti. It is to be inferred that this rodent constitutes the main food-supply of the kit fox in localities where both oeeur. The accompanying table shows the data appertaining to the speci- mens caught. Although a xerophilous species, examples from opposite sides of the river do not differ appreciably, in either cranial or external characters. The two females taken at Pilot Knob showed evidence of having recently suckled young, though this stage was well passed. LIST AND MEASUREMENTS IN MILLIMETERS OF VULPES MACROTIS ARSIPUS FROM THE COLORADO VALLEY Weight Tail Hind Mus. no. Sex Locality Date Ibs. Length vertebrae foot Ear 10581 2 Mellen, Ariz. side Feb. 26 21146 670 260 110 80 10582 @Q Ehrenberg, Ariz. side Mar. 27 3 725 290 120 85 10583 & Opposite Cibola, Calif. side Apr. 5 = 740 290 115 92 10584 Near Pilot Knob, Calif. side May 10 314 720 280 115 _ 10585 @Q Near Pilot Knob, Calif side May 15 3 724 297 108 86 Urocyon cinereoargenteus scotti Mearns Arizona Gray Fox The Colorado River bottom and nearby desert proved to be well populated with a form of gray fox. Specimens were taken, as shown in the accompanying table, on both sides of the river. Besides the localities represented by specimens, gray foxes were also seen on the California side near Pilot Knob; so that the species doubtless occurs continuously the whole length of the portion of the river explored by us. Our Colorado River material added a subspecies to the known mammal fauna of the state. Urocyon c. scotta is well distinguished from U. c. californicus, of the San Diego district, on the basis of [ Vou. 12 Zoology ications in Publ va ty of Californ iverst Un 256 9CIT 86S SFr OLT 9CL F009 Sr OZLT Ost 609 69% GALT TOL 069 GSh PLT Gol €6S Grr TALI Sol 6S Grr GALT 9CL §09 Shr FLT Gok 0109 Liar OT G&L 36s Gor T9T Tot PLS Sr SLT ° stip ° th 29 BES SO SES s Re Dion soins a=] D mR AS Boe ee oB Bt noe Sa ace 8 o4 ae 85 Be ose ai) tie tad 5 Ese Dae Eat eu3 @oe on Fso 2 3 S Ba ai A ees Bos _ oe omic 38 E ae *: ah cr] 5 ‘snyvoul JO YOJOU LOloJUB PUL SNtIejsod WNAdR, USWIRIOZ VOIMJod doULISTP WNUWIUTP 1 Lg9 G90 g'e9 380 Lt9 S60 OF9 SOL PFO = L'SOT 189 9°01 e789 ‘601 Le9 G90 S19 PLOT SF9 SGOT 48 8 es ne B 2 2 8" e at Bel at zr 18 98 avg O€T F6L O&L O&L LEL SéL 08T GEL GEL O&T a lt 068 066 COP CGP CIP OGF cer OIF SOF 4 =) 1S6 86 OF6 Cé6 OOOL oo uey 7 Ye OD ma T AAAnaA SH ct ‘ady OL We 6 ‘RW ¢g ‘RW g IW g ‘ABN g “Ie 13 “WT LE “49 oye OdBIOAYV “FURO ‘oyoworg “N “Tu 0% "yep ‘Aop[BA staontpourayy “yey ‘AopT[BA staonqamoyyD “FI[BD ‘Sa_pooN oy, aytsoddg “FeO ‘soe[paaN oy, eytsoddg “FBO ‘se[poan oy, eytsoddg “ZILY ‘SETPION OY, FO JOO ‘Z1ay ‘UeT[OW “FIO ‘sorpaon Aqpwoory AUTTIVA OGVYOTOD AHL WOU TL.L008 Of Ot “0 Of) 0/0) So 70) *O x” 2 n 6090T TO90T OO090T 66S0T 86S0L LZ6S0T 96S0L S6SOL F6SOL ‘ou “SHY SOQ LNDIIVOALANIO NOAIOOXYA AO SNAWIOUdS LINGVY JO SYALANITIIN NI SLNAWAYASVAW GNV LSIT 1914] Grinnell: Mammals and Birds of the Colorado Valley 257 both external and cranial characters. Scotti, as exemplified by the Colorado River series, has a longer tail, higher ear and paler colora- tion; the rufous tinges are less bright; the white endings of the hairs on sides of body and tail and on top of head are more extensive, giving a grizzled effect in the desert race. The skull of scoftti is of relatively lighter build; the teeth are smaller, being more slender and hence sharper; the auditory bullae are more inflated, that is, relatively higher and steeper-sided; the rostrum is conspicuously narrower. For measurements of californicus see Dixon (1910, p. 304) and for dis- cussion of relationships see Grinnell and Swarth (1913, p. 373). Mephitis estor Merriam Arizona Striped Skunk A common species in the riparian strips along both sides of the river. Data pertaining to the eight specimens obtained is given in the accompanying table. This animal evidently does not stray out on to the open desert, being probably kept to the bottomlands by daily need for water; nor has it the physical ability to cover much distance in a night’s travel. Those caught were in mesquite, screwbean, willow or arrowweed tracts. Tracks were often seen on the mud at the water’s edge. A male, no. 10576, weighed three pounds, two ounces. The female taken near Pilot Knob May 12 contained four foetuses. This species of Mephitis is distinctly smaller than VW. occidentalis. Although in all of the six skins of estor from the Colorado Valley there is much more white than in occidentalis, the amount varies indiv- idually to a large extent. In five out of the six there is a white pectoral patch, and in four of these there are additional fiecks of white mid-ventrally. The longer hairs of the tail are in each case fully white; but a thick growth of relatively short hairs clothing the tail is black terminally, white at base. As the season advances it looks as though the long white hairs of the tail tend to be shed, leaving the tail of a darker tone, until the black even predominates. The mid-frontal white streak varies in amount, but averages greater than in occidentalis. The dorsal white area (creamy white) is unbroken in no. 10575; in no. 10574 there are tufts of black in the mid-dorsal line over the rump; in the other four skins there is a well-defined black stripe separating the white into two lateral stripes LIST AND 10576 12648 10578 258 University of California Publications in Zoology | Vou. 12 which join on the shoulders, and continue forward to cover the whole back part of the head behind an abrupt line of demarcation joining the ears. (See pl. 13, fig. 21.) The question suggests itself whether or not geographic variation in proportion of white and black (as in Mephitis estor of the desert as compared with WM. occidentalis of the Pacific slope of California) is correlated with the maximum condition of efficiency from the stand- point of utility. Granted that the sole purpose of the skunk’s con- trasted markings is to offer a signal of warning, then the maximum of efficiency will be reached when the proportion of black and white is such as to bring the greatest degree of conspicuity amid the average of the natural surroundings. The surroundings on the desert, even in the riparian thickets, are far lighter-toned both night and day, than they are in the humid coast region. It would seem, therefore, that to secure the greatest con- spicuity the increased proportion of white must be provided for the darker surroundings, and conversely the increased amount of black must be presented in the lighter surroundings. This is the reverse of the case in fact. Hence it looks as though the warning theory does not gain substantial support from this direction. The problem of the significance of animal coloration promises important results when attacked upon a basis of the facts to be observed in geographic variation. COLORADO VALLEY Nature of Total Tail Hind Mus. no. Sex material Locality Date length vertebrae foot 10574 @ Skinand Calif. side, 5 mi. south Feb. 23 665 300 65 skeleton of Needles 10575 @Q Skinand Calif. side, Cheme- Mar. 9 625 300 70 skeleton huevis Valley ¢o Skinand = Arizona side, foot of Mar. 5 642 330 70 skull The Needles 10577 = Skullonly Arizona side, Ehrenberg Mar. 26 625 300 68 3 Skeleton Arizona side, Ehren- Mar. 24 610 288 71 only berg 6 Skinand Calif. side, opposite Apr. 6 525 300 70 skull Cibola Q Skinand Calif. side, 20 mi. N. Apr. 12 615 322 69 skull of Picacho 10580 @Q Skin and Calif. side, near May 12 595 280 64 skull Pilot Knob MEASUREMENTS IN MILLIMETERS OF MEPHITIS ESTOR FROM THE 20 22 1914] Grinnell: Mammals and Birds of the Colorado Valley 259 Spilogale arizonae arizonae Mearns Arizona Spotted Skunk We were told of the presence of ‘‘hydrophobia’’ skunks among The Needles, Arizona. Mr. Stephens saw tracks of one in the river bottom on the California side five miles below Needles. The species is certainly not common, or our traps would have given more indica- tion of its presence. The only specimen procured by our party was trapped in the arrowweed belt within one hundred yards of the river, on the California side at our last station, close to Pilot Knob. It is an adult male (no. 10573) with following external measurements by collector (Stephens) : length 440, tail vertebrae 170, hind foot 46, ear 22, weight 18 oz. The cranium measures: basilar length 49.8, zygomatic breadth 36.1, mastoid breadth 30.9, interorbital width 14-4, height of brain-case (as measured by Howell, 1906, p. 37) 17.8. I sent this skin-with-skull to the Bureau of Biological Survey, Washington, where Mr. A. H. Howell determined it to be ‘“‘arizonae but not typical.’’ He further considered it ‘‘as grading toward martirensis,’? which, from its geographical location, is quite to be expected (see Howell, 1906, pp. 29, 30, pl. 1). The present record is the basis of the first assignment of arizonae to Californian territory. Taxidea taxus berlandieri Baird Mexican Badger While no badgers were secured by our party directly, I purchased a flat skin, with skull, of a young one at the Barber Ranch twenty miles above Picacho. This had been killed on the California side in the autumn of 1909. Badgers were reported as frequently seen in the vicinity of the Draper ranch, eighteen miles north of Picacho, and also on the Arizona side in the neighborhood of Cibola. The skin secured (no. 10603) as compared with one of about the same age from Ventura County, California (no. 7078), is very much paler. The individual hairs have longer white tips, and a mid-dorsal white stripe extends from top of nose to rump. In the Ventura County specimen there is a white stripe from tip of nose to between shoulders, only. The skull of no. 10603 is smaller, doubtless due to younger age, but it appears to be also of relatively lighter build. 260 University of California Publications in Zoology — [Vou. 12 Procyon pallidus Merriam Pallid Coon An abundant representative of the riparian association everywhere. The accompanying list shows loealities of capture from Needles to Pilot Knob, and also from one of the Colorado’s distributaries, New River, in the delta region near Imperial, California. At practically every place where we had oceasion to examine the muddy margin of the river or its lateral sloughs, the conspicuous foot- prints of coons were to be seen. One coon was caught in an unbaited trap set for beaver beneath the surface of the water, at the river’s edge. The chief food of the coons was evidently fish. At a drying-up overfiow pond, the shallow water of which was crowded with catfish, a profusion of the tracks of herons and eoons showed a marked com- munity of interests on the part of these two fishers. No young-of-the-vear came to notice. The female caught May 11 near Pilot Knob contained four embryos. The one caught near Imperial May 10 contained five embryos. The weight of the female no. 10610, was 16 pounds. Two fat males, nos. 10606, 10607, weighed 1814 and 18 pounds respectively: A lean male, no. 10609, weighed but 13 pounds. The very large old male (no. 7153), from near Imperial, weighed 20 pounds. In the original description of Procyon pallidus (Merriam, 1900, p. 151), both external and cranial characters of this form are given. The cranial peculiarities mentioned, as compared with Procyon psora, do not appear to hold in our material. There is much variation in the respects mentioned, as will be understood from the table of measure- ments. In coloration alone, however, there is abundant basis for the recognition of pallidus. The seven skins from the Colorado Valley, ineluding Imperial Valley, although somewhat variable in coloration, are in mass effect conspicuously different from the coons of the Pacifie Coast region. A large series of skins of the latter are available for comparison; and even selecting extremes from the pallidus and psora-pacifica series, no overlapping of characters can be found in the material at hand. The general pallor of pallidus as compared with psora is due to: (1) The far less amount of black on the individual hairs everywhere except on the nose and cheeks. This is the converse of saying that much more of each hair is white. Along the sides. on the ears and on the light intervals between the tail-rings the hairs are pure white with no black tippings. (2) The black areas are less in extent. The 1914] Grinnell: Mammals and Birds of the Colorado Valley 261 seven or eight black tail bars are narrower, thus making the light intervals broader; the whole tail is thus very much lighter colored. (3) The ‘‘black’’ color of the hairs is in many places not so intense, being, particularly on the tail and head, of a deep vandyke brown. Fading evidently brings a browning of the dark hairs to some extent. (4) The under fur is very much paler in color. Dorsally it is vandyke brown instead of bistre; on the sides and belly it grades through cin- namon to pale clay color. In the facial region the typical coon pattern is present, but the white markings are not so conspicuously contrasted. This is due to the less intense black, and to the fact that the pileum is seareely darker than the rest of the back. The post-mental area and the median nasal stripe are rather pale vandyke brown. Only the transocular black patch is retained in nearly its usual distinctness. The whiteness of the backs of the ears is a notable feature; and there is altogether lacking the black patch behind the ears usually so well marked in psora. In a general way pallidus thus differs from psora in very much paler coloration. The two specimens from the vicinity of Imperial are topotypes of pallidus, and accord closely with the brief description of the form as given by Merriam. These two topotypes are the palest of our pallidus, but it will be observed from the list that they were secured in May. As might be expected from this circumstance, they had evidently been subjected to much more wear than especially our Needles and Mellen specimens, which are somewhat darker. It is easy to see that the cutting off of the black tips of the long hairs in the latter, particularly on the mid-dorsal region, would result in a decided paling in the general color. Eliminating this factor of wear, and the slight amount of fading which very probably oceurs (even though these nocturnal animals may not be exposed to intense sunshine to any large extent as they probably spend the day in drift piles or thickets rather than in trees), I cannot see that those specimens from farthest up the river are any paler than those from farther down the river, or than the topotypes. It would not be reasonable, therefore, to hold that the apparently darker animals farthest up the river from the assumed center of differ- entiation (Colorado delta) indieate intergradation with darker coons to the northward, even though this might be expected. It is even pos- sible that the cliff-confined river above forms an effectual barrier to the north, so that there has been little chance of mingling of coon strains from elsewhere. Pallidus may be hemmed in by ordinarily uncrossable barriers, and hence has attained a clear-cut specific identity. [ Vou. 12 University of California Publications in Zoology 262 OL 16 Wqio WIweUeq yesaf yo WPL BOL Lot giv Piyenjestin 0'S9 o'P9 OOP Ly G'8P O'Ob oe wu [eweyed-3s0g 3Su09 [S31qsora3Uy TOILE S96 eq i ore s 28 WINIySAI JO YIP! joa eulueo oR yred 3a “HOM podxs, OL6L OF UO pongo you oy ‘HUOydoIg WE Aq GOBL Ut POVOT[O) 1 9 POL 0'90L GILT oUt BLU Pesuex Jo wIsugq sepseg — oY i) [L vg O61 VITO UNV 008 068 O18 O08 BAG S08 =| ra) OLS 11 An gg ‘ady i8 Av! OL AnpW owed qousy Voptg ‘ddo ‘opia ‘ziay FEUD HOLONIO FUUD ‘Oyouorg JON ‘Tu Og AIL MOL [O[Ny JIB) ‘Ho[poonN fo "gy "run g UD ‘Tenwodury yo tay “cur g FUUD ‘Turtodury yo ‘ay ‘rar g AVoory ‘pu po ‘pu dd’ ‘pu dd "pu plo ‘pu pro "pu po xOR OLD0T BODO L HODOL LODO HO90T ‘ou WA, AUTIVA OG VHOTOD WHHL WOU SAd1TIVd NOAQDONd AO SUSLLOWITITIN NI SLNSUNCOLOS VOL CNV LSI 1914] Grinnell: Mammals and Birds of the Colorado Valley 263 Corynorhinus macrotis pallescens Miller Pale Lump-nosed Bat At Riverside Mountain, California, March 18, three bats were found at the end of a sloping drift in the Steece copper mine. They were clinging to the rock wall, and at once took flight, attempting to pass us towards the opening of the tunnel. Two were secured, nos. 10694, 10695. The fur of these has a slight reddish east, which is doubtless wholly adventitious, due to the fine, sticky red dust with which the walls of the mine were covered. Both were females, one of them containing a single embryo. Antrozous pallidus pallidus (LeConte ) Desert Pallid Bat On several nights in April, large light-colored bats were moment- arily observed flitting about in the moonlight close over our beds. At times the flutter of their wines was clearly audible. But they seldom appeared until it was too dark to shoot. On April 20, on the California side, eight miles east of Picacho, a specimen was shot at late dusk as it flew among willows across a patch of open sky. This, the only example of the species obtained (no. 10696), was a female containing two embryos. Myotis occultus Hollister Hollister Bat This very distinet species was only recently described (Hollister, 1909, p. 43) from two specimens taken May 14 and 15, 1905, on the California side of the river in the bottomlands ten miles above Needles. The fact that our own expedition failed to detect the presence of this bat until the first week in May would point towards its late spring arrival in the region generally. We obtained six specimens, as listed in the accompanying table of measurements. The first was shot at late dusk close to the river bank between files of cottonwoods, in just the same association as those taken by Hollister. At our second locality of capture, the remaining five specimens were shot over the water in a back eddy of the river. Here these bats arrived in considerable num- bers at early dusk to drink, flitting down to the water’s surface and dipping several times before flying off among the willows and cotton- woods. We used a boat in shooting and retrieving the specimens obtained. [ VoL. 12 319 15 15 15 ao w~ novo a id un wD Ww AIBTIXB]T 19 University of California Publications in Zoology MOJ-4300} 00 264 OF 92 86 UST 9°GE 9°¢ Lg Loe. €18 S}[NPB XIs oy} JO osvIOAY OF Lh 86 OST ce FS 6 8& 148 p Av “FUR “Bun A FO "ON “Tu G 4 2L0L01 OF 92 9°6 €St £96 as 8 LE 148 g Avy ‘Fey ‘vung JON THUG =P g90LOT UP eL 9°6 SPL OSE v9 Gs PE 98 p Av “JUV “VuUIN A FO “GN “Tut G 6 GOLOL OP OL TOt ¥vST P9E 09 6 96 06 ¢ Av “FUBH “VUINA FO “GEN “TG & POL0L TP i Li Ci ccs e's 68 iad 68 ¢ Av “FB “BUN A FO "GEN “TU G & g0L0L 68 Lh 86 ol PGE 6S 6 Gs G8 T Av “FILED ‘Sopoyypo JO *g “TU pF & Z0L0T OH ow oN ook hy 8 boo Bar| a] awa Aqywoory xog “ou ‘sn Be sa ga a86 & B 8s 482 &¢ fo Be 88 wes = EI ~ ¢ esa ee asso eh 9 Bin FI 2 a oe 6° S ao © Be tes AQTITVA OGVUOTOD AHL WON SALTAI00 STLOAW AO SUALAWITTIIN NT SUNG OS Vo 1914] Grinnell: Mammals and Birds of the Colorado Valley 265 I sent one example to the Bureau of Biological Survey, where Mr. A. H. Howell made the specific determination here employed. Our series bears out closely the characters assigned in the original descrip- tion (Hollister, 1909, p. 48), and they are in all these respects sur- prisingly uniform. The very broad and flat-topped rostrum and braincase constitutes a character for discrimination from all other species of Myotis in California except orinomus. The proportions generally are peculiar (see table of measurements). Fig. H. Right upper dental series Fig. I. Right upper dental series of Myotis occultus, no. 10702, 9; mid- of Myotis occultus, no. 10706, g; mid- dle upper premolar absent. X 4. dle upper premolar present. X 4. Variation in general proportions from those of preceding figure possibly due to age. An interesting fact pointed out by Hollister is the variability in a feature usually considered of much more fundamental importance than the external characters employed in distinguishing members of the genus, namely, the presence or absence of the middle upper pre- molar (pm*). In one of Hollister’s specimens this tooth was present, in the other absent; in three of ours it is present, in three it is wanting. Thus fifty per cent of the individuals so far collected lack the tooth in question, certainly a remarkable aberrancy from the norm in the genus Myotis (see Miller, 1907, p. 201), and denoting a tendency to specialization in this member of the genus, along a line regularly shown in other closely related genera (see figs. H, I). Myotis californicus pallidus Stephens Stephens Little Pallid Bat The accompanying table shows certain circumstances of capture of this species. Although obtained at but the two loealities, Mellen, on the Arizona side, and opposite The Needles, on the California side, 266 University of California Publications in Zoology — [Vou. 12 I was fairly sure that I saw the same species at other localities along down the river. Those obtained were all shot at late dusk, considerably later in the evening than most of the appearances of Pipistrellus hesperus. Instead of flying high, against the sky, as in the ease of the latter species, MW. c. pallidus was almost always foraging low over the bushes of the second bottom, or along shallow washes between clumps of mesquite, seldom appearing above the sky-line. The move- ments of flight were peculiar also. LIST AND MEASUREMENTS IN MILLIMETERS OF MYOTIS CALIFORNICUS PALLIDUS TAKEN IN 1910 ON THE COLORADO RIVER Tail Hind Mus. no. Sex Locality Date Length vertebrae foot 10698 é Mellen, Ariz. Feb. 26 75 38 5 10699 Jo Mellen, Ariz. Feb. 28 77 38 6 10700 é Opposite The Needles, Calif. Mar. 1 81 40 6 10701 2 Opposite The Needles, Calif. Mar. 3 75 35 6 The four specimens obtained are uniform among themselves and with a topotype specimen of Myotis californicus pallidus (no. 7350) from Vallecito, on the western side of the Colorado desert in extreme eastern San Diego County. All agree closely with the description of M. c. pallidus (Stephens, 1900, p. 153). An additional feature, as compared with Myotis californicus californicus from Monterey, California, is the smaller skull of pallidus, with decidedly smaller brainease, less inflated in the parietal region. Myotis velifer (J. A. Allen) Cave Bat Not obtained by our party; but there are in the Museum three skins- with-skulls (nos. 7762-7764) taken by Charles Camp at Needles July 16 and 18, 1909. Mr. Camp states that this species was roosting in numbers in an old storehouse from which they were routed out and shot. One of the specimens was forwarded to the Bureau of Biological Survey, Washington, where the above determination was confirmed by A. H. Howell. I do not find a previously recorded oceurrence of this species for California. 1914] Grinnell: Mammals and Birds of the Colorado Valley 267 Pipistrellus hesperus hesperus (HH. Allen) Canon Bat The most abundant representative of the order Chiroptera observed during the period of our work. Seen abroad at dusk as early in the season as February 23, when the nights were still so cold that ice formed in suitable places. Numerous at Mellen, February 23 to 28, and swarming in the vicinity of The Needles March 1 to 3. Thence- forth seen at nearly every station all the way down the river. One thing was conspicuously noticeable in regard to occurrence, namely, that this bat varied directly in degree of abundance with nearness to cliffs, or hillsides with outcroppings of fractured rock. In other words, this species probably dwelt exclusively in the rocks during the day, from which it emerged at early dusk to forage out over the river bottom in the near vicinity. Individuals were often seen before the sunlight had yet left the eastern hilltops. On one occasion, as we were floating down the river near Pieacho, a Pipistrellus appeared in flight in the glaring forenoon sunshine, dipped down to the surface of the water, where it touched, and thence flitted back to a crevice in the nearby cliff. Seventy-four specimens of this species were shot (nos. 10382-10423, 10746-10777), 42 being preserved as skins and 32 as alcoholics. Kighteen out of this series are from the California side of the river near Pilot Knob, only about six miles due west of old Fort Yuma, and are thus practically topotypes of Pipistrellus hesperus. In reviewing the Museum’s entire collection of Pipistrellus from California, it becomes clearly apparent that while there is but one species represented, there are two appreciably different subspecies of this species, one a pale-colored form occupying the arid desert regions from Owens Valley and the vicinity of Walker Pass southeast to the Mexican line, the other a darker-colored form occurring on the Pacifie slope of southern California (in the San Diegan district) and in certain parts of the San Joaquin and Sacramento valleys. It appears to the writer that we have here two races well worthy of recognition in nomenclature. The name hesperus was based on the desert form. A name is apparently available for the Pacifie race in the Vesperugo merriami of Dobson (1886, p. 124). This has always been synonymized under Pipistrellus hesperus. While the habitat is given by Dobson as ‘‘North America (Locust Grove, State of New York) ’’, an error was committed, for the type really came from Red Bluff, Tehama County, California (fide Miller, 1897, p. 31). Dobson 268 University of California Publications in Zoology [Vo. 12 received his specimen from C. H. Merriam, whose residence at that time was Locust Grove, New York. The case seems to be clear, and I propose that the Pacific slope race be called Pipistrellus hesperus merriami (Dobson), the type locality of which is thus Red Bluff, California. The characters of this form, as compared with P. h. hesperus, lie in the darker, distinctly browner tone of coloration both above and below, and in somewhat larger size throughout. The Museum has specimens, unequivocally of this form, from the following localities, all within the state of California: Marys- ville Buttes, Sutter County ; Raymond, Madera County; Cuyama Val- ley, Santa Barbara County; Fort Tejon, Kern County; San Fran- cisquito Canon, northern Los Angeles County; vicinity of Pasadena; Escondido, San Diego County. Eptesicus fuscus (Beauvois) Large Brown Bat One shot at dusk on the California side near Pilot Knob, May 6. Other bats supposed to be the same species were seen flying down the river high overhead the same evening. A strong west wind was blowing at the time. The species was not seen at any other place. The specimen secured (no. 10697) is an adult female. It appears to differ in small size and extreme paleness from the average of the species from California. It about equals in the latter respect the palest out of a series of ninety brown bats from the Pacifie slope of California. The color dorsally is uniform isabella color, ventrally pale wood brown. Measurements: length 107 mm., tail vertebrae 44, foot 9, forearm 42.5, longest finger 72, ear (dry) 12. The skull, too, is appre- ciably small. A general inspection of the Museum’s series of this species from California points towards the existence within the state of at least three geographic races based on size and depth of color. But so much of the total area is unrepresented by specimens that systematic analysis at this time seems inadvisable. Nyctinomus mexicanus Saussure Mexican Free-tailed Bat We were fairly certain of seeing this bat at almost every station, as a rule flying high and often squeaking loudly. Only three specimens were secured (nos. 10690-10692) the first two at Mellen, February 26, the third in Chemehuevis Valley, March 11. 1914] Grinnell: Mammals and Birds of the Colorado Valley 2 for) wo Macrotus californicus Baird California Leaf-nosed Bat One of the rooms of an adobe ruin on the Arizona side, ten miles below Cibola, showed considerable bat excrement on the floor. No bats could be found in the thatch above, so the place was visited in the evening. No bats appeared in the vicinity until all daylight had faded. Then two were seen flying about the ruins; and by means of a lantern and butterfly net, one was caught in the room referred to. This was probably used by the bats merely to repair to while eating the moths caught outside; wings of the latter were strewn upon the floor. The specimen obtained (no. 10693) is an adult male; the date April 8. Transmitted June 13, 1913. LITERATURE CITED ADAMS, C. C. 1905. The postglacial dispersal of the North American biota. Biol. Bull., 9, 53-71. AMERICAN ORNITHOLOGISTS’ UNION COMMITTEE, J. A. ALLEN, Chairman and Editor. 1910, Check-list of North American birds. Ed. 3, revised (New York, American Ornithologists’ Union), 430 pp., 2 maps. BaltLey, V. 1902. Synopsis of the North American species of Sigmodon. Proe. Biol. Soe. Wash., 15, 101-116. Bairp, S. F. 1861. ‘‘List of Birds collected on the Colorado Expedition’’ in ‘‘Zoology’’ in Ives, Report upon the Colorado River of the West, pp. 5-6 (of Part V). BENDIRE, C. E. 1892. Life-histories of North American birds with special reference to their breeding habits and eggs. U. 8. Nation. Mus., special bulletin, 1, pp. vili+446, pls. 12. BRENINGER, G. F. 1898. Hybridization of flickers. Osprey, 3, 13. BREWSTER, W. 1888. On a collection of birds lately made by Mr. F. Stephens in Arizona. Bull. Nutt. Orn. Club, 8, 21-36. 1887. Three new forms of North American birds. Auk, 4, 145-149. Brown, H. 1908. Arizona bird notes. Auk, 20, 48-50. 1904. The elf owl in California. Condor, 6, 45-47. 1906. The water turkey and tree ducks near Tucson, Arizona. Auk, 23, 217, 218. 270 University of California Publications in Zoology — | Vou. 12 Cooprr, J. G. 1861. New California animals. Proce. Calif. Acad. Sei., 2, 118-123. 1868. Some recent additions to the fauna of California. Proe. Calif. Acad. Sci., 4, 3-13. 1869. The naturalist in California. Amer. Nat., 3, 470-481. 1870a, The fauna of California and its geographical distribution. Proe. Calif. Acad. Sci., 4, 61-81. 1870b. Ornithology of California. Volume I. Land birds. Edited by 8. F. Baird, from the Manuseript and Notes of J. G. Cooper, xi+592, many text figs. Cougs, E. 1866a. From Arizona to the Paeifie. Ibis (2), 2, 259-275. 1866b. List of the birds of Fort Whipple, Arizona; with which are incor- porated all other species ascertained to inhabit the Territory; with brief critical and field notes, descriptions of new species, ete. Proce. Aead. Nat. Sei., Phila., 1866, 39-100. 1903. Key to North American birds. Fifth edition, 1, xli+1—535 pp., 2, vi+537-1152 pp. Court, E. J. 1908. Treganza Blue Heron. Auk, 25, 291-296. Daccert, F. 8. 1902. Winter observations on the Colorado desert. Condor, 4, 37-39. Drxon, J. 1910. A northern coast form of the California gray fox. Univ. Calif. Publ. Zool., 5, 303-305. Dosson, G. E. 1886. Description of a new species of Vesperugo from North America. Ann. Mag. Nat. Hist., (5), 18, 124-125. Dwieut, J., JR. 1900. The sequence of plumages and moults of the passerine birds of New York. Ann. N. Y. Acad. Sci., 13, 73-360, pls. 1-7. Evuiot, D. G. 1905.