caren sehen ate ining tonto a - a Ms SS yememen nee Reet rer ‘ " at 2 ney pe ae te! : : ies i) Rhy, Ge iy ( a oe ee] cs ‘ a : , eye ; | ' r . =, 5 } ; 7 : : ? ah: ty | , : a \ : So SPR EEDING SC | cpp “rve7 @ maomiorna Acadent (OL Screncs wen vie . f —< 5 Le ee see 5 4 | , . ° ‘ P ie. ‘ ; ‘ . ’ * y y . } a * ea - ‘ ‘ - + ) Houotryee: A dead but well-preserved shell with general characters as de- seribed for the genus. Total whorls six and three-quarters, consisting of about one and one-half nuclear turns and five and one-half postnuclear turns. The embryonic tip is small but enlarges rapidly during the second half-turn, so that this portion of the nuclear whorls projects shehtly above the plane of the first two postnuclear whorls. Spire very slightly concave except for the projecting nuclear area. Sutures well impressed. Umbilicus very wide and shallow. Maximum diameter, 6.8 mm.; minimum diameter, 6.4 mm.; height, 1.6 mm. Holotype in the California Academy of Sciences (Paleo. Type Coll., 10289), collected in Violin Cave on South Fork of Dry Creek, west bank, Amador County, California (See. 4, T7N, R12E, MDM, USGS Pine Grove Quad.), by Edward Danehy and Arthur Lange, November 22, 1951, Stan- ford Grotto, 142. PARATYPES: 33 specimens collected with the holotype, of which about 20 are full-grown though all are dead shells in a more or less bleached condi- tion. Paratypes are in the California Academy of Sciences (Paleo. Type Coll., 10290, 10291, and 10292), Stanford University, the Los Angeles Mu- seum, the San Diego Society of Natural History, the Academy of Natural Seiences of Philadelphia, the United States National Museum, and the pri- vate collections of Dr. S. S. Berry, Dr. W. O. Gregg, Walter Miller, M. L. Walton, and Leshe Hubricht. OTHER Locauity: Black Chasm, near Voleano, Amador County, California, on the east side of the south branch of Sutter Creek, at an elevation of 2040 feet. Two lots were taken: the first containing nine bleached shells collected by Charles Moller, January 138, 1951 (C.A.S. 34824); the second, ten shells collected by Harold Treacy, May 5, 1951 (C.A.S. 34825). All were dead shells found deep down in this almost vertical shaft-like cave. ReMARKS: This remarkable small, closely-coiled, flat snail looks like a large Helicodiscus but the lack of internal teeth and the absence of even the slightest vestige of spiral sculpture immediately set it apart from this genus. Similar snail shells from the San Francisco Bay region have been mentioned in the literature by W. G. Binney (1886; 1890), Hanna (1939), and Pilsbry (1948) under the names Helicodiscus lineatus Say, H. fimbriatus var. sal- monensis W. G. Binney, and H. salmonaceus W. G. Binney. Although speci- mens from these early collectings have been at hand for many years, the lack of sufficient specific information on them and the fact that they have not been taken by any collector since has caused hesitation in recognizing VoL. XXIX] A. G. SMITH: SNAILS FROM CALIFORNIA CAVES 35 them as part of the snail fauna of California. Discovery of these snails in fair quantity in two Amador County limestone caves leaves no doubt about their authenticity as well as their novelty. Although not one was found alive, several specimens in the type lot are fresh enough to lead one to the supposition that living snails of this species may be found by diligent search in the vicinity of the caves where the dead shells were collected. Discussion: Violin Cave is about eight miles southeast of the old gold set- tlement of FiddJetown (hence the name of the cave) and is around three or four miles air-line from Voleano and the Black Chasm location. Fiddler’s Cave is nearby. The cave entrance is 30 feet above the South Fork of Dry Creek at the base of a massive marble outcrop. Mr. Danehy has furnished the following comments: Violin Cave was almost dry when visited in November, 1951, although in the low, narrow crawlway where the estivating snails were found the gravel floor was damp. This point is approximately 50 feet from the entrance and was the beginning of the zone of total darkness. The cave is nearly hori- zontal and the farthest point reached is estimated as being 250 to 300 feet from the entrance. Here, soil and pebbles from the surface are accumulated below crevices. ... It was in the crevices and upon the soil that the dead shells were found. Almost every cave I have seen has places where soil and debris come in from the surface, even though no openings are evident. The snails taken from this cave were selected as the type lot because of the number of specimens collected, which permits a liberal distribution of paratypes. Measurements of the ten largest specimens are as follows: Dimension Range Average Maximum diameter .............. 6.0 — 7.4mm. 6.6 mm. ATT yp ee a eR os one de e 1.4 — 1.6 mm. 1.5 mm. Number of whorls ................ 64-7 634 All the shells in the lot show marked similarity in characters and seulp- ture. They differ only in shght irregularities in the coiling of the whorls. Black Chasm, where Speleodiscoides was first rediscovered, is located on a hilltop in an area of massive outerops of marble, weathered smooth into fantastic shapes in places—a characteristic of several locations in the Mother Lode country. It is described as follows in the January, 1951, re- port of the Stanford Grotto: It is an L-shaped vertical fissure in limestone, first entered by E. Sammis with a party in 1854. At that time the opening was about 18” across but has since been enlarged to a diameter of several feet in subsequent attempts to commercialize the cavern. The entrance confronts a steeply inclined mud slope composed of material washed in from above. Several vertical pitches follow before the first landing is reached—an estimated 120 feet below. From the entrance down to this landing there are found no foot- or hand-holds. 36 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. Fifty or sixty feet below the landing lies the bottom of the cavern—a series of interconnecting (?) lakes, the first of which has a length of about 70 feet. Depths were estimated to exceed 20 feet. Between the landing and the lake level occurs one mud platform. The entire downward route is mud-coated, but away from the range of debris the cavern walls appear unsullied. The snail shells were reported as being found in a side passage extend- ing off from and above the first landing about 120 feet down from the mouth. How they found their way into this side passage is hard to imagine unless it has had some sort of surface connection. The Black Chasm locality was visited by the writer in May, 1952. A fairly thorough search of likely-looking piles of marble and country rock in the vicinity was made on the theory that the snails in the cave had dropped or been washed in from the surface. No snails were found, however, either of this or any other species, possibly owing to the fact that over much of the entire area were evidences of surface alteration resulting from placer or other gold-seeking operations. Time was not available to devote a simi- lar search around Violin Cave. The freshness of specimens from this loca- tion would appear to make such a search more likely to be successful if the species is still living. Shells of Speleodiscoides from the San Francisco Bay region that are available for comparison consist of two from Sausalito from the Reverend Joseph Rowell Collection, and two others from the W. J. Raymond Col- lection. Of the latter, one was labeled as collected at Sausalito by Voy in 1872, the other being found by Professor Raymond himself at “Laundry Farm” in the vicinity of what is now Leona Heights, a suburb of Oakland. All these specimens were presumed to be fossil or at least semifossil, a view that now seems questionable in the light of recent finds. Two of the four shells are adult or nearly so. They do not have a fossil appearance. Of the other two, one is young; the second is badly worn and broken. Two of the good specimens have the following measurements : CAS No. AGS No. Locality Collection Max. Diam. Height No. Whorls 34826 1937 Sausalito, Calif. Rowell 5.5mm. 1.5mm. 5% 34827 8811 ? Raymond 6.1mm. 1.4mm. 536 The two Raymond specimens were together in the same bottle, making it impossible to tell which one Voy collected and which Raymond himself found. The principal account of these snails was furnished by J. G. Cooper and published in an early Bulletin of the California Academy of Sciences (Cooper, 1885). As copies of this Bulletin are not readily accessible, it is pertinent to this discussion to quote what Cooper has to say: The fourth species (or specimens) of discoid sub-fossils have been in my Vou. XXIX] A. G. SMITH: SNAILS FROM CALIFORNIA CAVES 37 hands since 1871, when Mr. C. D. Voy brought them to me as discovered by him “in an Indian mound north of S. F. Bay.” They are of the size and nearly the form of Helicodiscus lineatus (a species said to occur in California), but without a trace of teeth or sculpture. I considered them, however, imperfect specimens of that form until 1883, when W. O. Emerson found a few of the same much larger but more fossilized, near Haywards, on a sliding bank, where they seemed to come from under the roots of a large tree that had been undermined. These also are toothless, and some of both sent to Mr. Bland and Mr. Binney, have been decided to be a new species if not a new genus. As with the specimens of Hyalina, mentioned before, these were where they could not have been washed down by mountain streams. Without living specimens it will be inadvisable to describe these, especially as their being found in an Indian mound is some indication that they may increase instead of decrease with the settlement of the country. The chances, however, seem to be in favor of their being cave dwellers or more northern species, introduced where found by birds, but not able to increase there. This account is interesting from the information it gives about the con- ditions under which these snails were found. But it is all the more notable because of Cooper’s supposition that they might be cave-dwellers. It is difficult to see the reasoning back of this unless Cooper, who was a keen observer, had information in addition to what he published. Whether it was a shrewd guess or was based on more tangible evidence, the fact remains it has been proved true after a span of nearly seventy years! Compared with specimens from the two caves in Amador County, the San Francisco Bay shells have a flat rather than a slightly coneave spire, less prominent transverse rib-sculpture, and the whorls appear to be not quite as tightly coiled. As these differences are so slight, and as only two good specimens are available for comparison, determination as to whether they constitute a recognizable geographic subspecies is left pending the col- lection of additional material from the San Francisco Bay area. List OF CALIFORNIA CAVES EXPLORED, WITH RECORDS OF THE SNAIL SPECIES COLLECTED SHasta County* 1. SAMWEL Cave. Sec. 5, T35N, R3W, MDM, Shasta Nat. Forest Map; Stanford Grotto, 23; Edward Danehy eoll., September 20, 1950. Picked up on the ground on a prominent limestone bluff known as Eagle Rock, several hundred yards west of the cave entrance. Monadenia troglodytes Hanna & Smith. Two bleached immature specimens. Discovery of this species outside of the cave indicates that it may be found living in the vicinity, though described origi- * The counties are cited in geographic order from north to south. 38 bo CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. nally as a Pleistocene fossil. The depressed spire of the specimens, though quite immature, gives credence to the correctness of the identification. PLuMAS CoUNTY JUNIPER CAvE. Sec. 17, T23N, RIOE, MDM, USGS Downieville Quad.; Western Speleological Inst., C—96, C-103, and C-105; Raymond de- Saussure coll., July 20, 1953. Taken on the floor and under rocks in the Cavern Room in the zone of total darkness. Helminthoglypta proles (Hemphill). One adult and one broken specimen; dead. Pristiloma juniperum A. G. Smith, new species. Four specimens; dead. AMADOR COUNTY Buack CHASM. 34 mi. south of Voleano, Stanford Grotto, 68 and 104; Charles Moller and Harold Treacy ecolls., January 13 and May 5, 1951. “Found in side passage about 120 feet from the surface, in total dark- ness.”’ Speleodiscoides spirellum A. G. Smith, new genus and species. A total of 19 dead specimens, a few with deposits of small calcite erystals on one surface. Pristiloma subrupicola spelaewm (Dall). A single dead specimen. ViouIn Cave. Sec. 4, T7N, RIZE, MDM, USGS Pine Grove Quad.; Stanford Grotto, 142, Edward Danehy and Arthur Lange eolls., No- vember 25, 1951. Named for its location near Fiddletown. Speleodiscoides spirellum A. G. Smith. Thirty-four dead speci- mens. The type lot. Pristiloma subrupicola spelaeum (Dall). Two living and two dead specimens. RIppLeD Cave. Sec. 28, T7N, RI1E, MDM, USGS Big Trees Quad.; Stanford Grotto, 144 and 145, Gil Lang eoll., February 3, 1952. Helninthoglypta cypreophila (Binney & Bland). One living small adult. Pristiloma subrupicola spelaeum (Dall). Three living specimens. HarE Cave. Sec. 31, T6N, RI1E, MDM, USGS Sutter Creek Quad.; VoL. XXIX] A. G. SMITH: SNAILS FROM CALIFORNIA CAVES 39 10. 188 Western Spel. Inst., C-34, Raymond deSaussure ecoll., April 5, 1953. Taken near second entrance in the twilight zone. Helminthoglypta cypreophila (Binney & Bland). Three dead ju- veniles. SUTTER CREEK CAvE. Sec. 29, T7N, R12E, MDM, USGS Pine Grove Quad.; Western Spel. Inst., C-61, Raymond deSaussure coll., May 25, 1953. Collected in lower right passage below a fissure, on the floor in total darkness. Haplotrema alameda Pilsbry. Twenty-seven adult and three young specimens, all dead. ELDORADO COUNTY PIONEER CAVE. Sec. 18, T12N, R9E, MDM, USGS Auburn Quad.; Stanford Grotto, 31, Edward Danehy ecoll., October 7, 1950. “Cave open to daylight due to large entrance.” Monadenia mormonum (Pfeiffer). Collected in 1875 by J. G. Cooper but not found there by the Stanford speleologists. Ammonitella yatest J. G. Cooper. One dead adult specimen. CALAVERAS CoUNTY CAVE OF THE CATACOMBS (MILLER’S CAVE). Sec. 2, T4N, RI3E, MDM, USGS San Andreas Quad.; Stanford Grotto, 139, Edward Danehy eoll., Mareh 28, 1951, and November 23, 1951. “Taken in the zones of twilight and total darkness. Some appeared washed [in] and others 5 seemed to be estivating.”’ Monadenia mormonum (Pfeiffer). Six living adults; eight adult and nine immature specimens, all dead. Helminthoglypta cypreophila (Binney & Bland). Two adults and one immature specimen, all dead. Ammonitella yatesi J. G. Cooper. Forty-three adults and 11 im- mature specimens, all dead. CAVE OF THE SKULLS. North side of the Stanislaus River, about half way between Columbia and Vallecito; Stanford Grotto, 133, Edward aa Danehy coll., Mareh 25, 1951. “Dead shells taken in the twilight zone.” Monadenia mormonum (Pfeiffer). One dead adult shell. Cave City Cave. Sec. 14, T4N, RI3E, MDM, USGS San Andreas Quad.; Stanford Grotto, 9a and 9b. Some shells taken January 7, 1950, 40 13. 14. CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. in total darkness several hundred feet from the nearest of the three cav- ern entrances. Others collected March 24, 1951, about 40 feet from the entrance in the twilight zone; Edward Danehy coll. Also, Western Spel. Inst., C-173 and C—174, several hundred feet into the cave in total darkness, probably washed in; Raymond deSaussure, coll., February 7, 1954. Monadenia mormonum (Pfeiffer). Several dead specimens. Also collected by Hanna & Rixford (1923). Helminthoglypta cypreophila (Binney & Bland). One dead adult. Ammonitella yatesi (J. G. Cooper). Original lot found by L. G. Yates about 100 feet within the mouth of the eave (Cooper, 1868). Not reported by the Stanford Grotto but later a series of 60 speci- mens (topotypes) were collected by Raymond deSaussure of the Western Speleological Institute. Pristiloma subrupicola spelaeum (Dall). Thirteen dead specimens (topotypes). SHaw’s Cave. Near Mercer’s Cave; Stanford Grotto, 36, Edward Danehy coll., October 8, 1950. “Taken in total darkness. Washed in by water.” Monadenia mormonum (Pfeiffer). One dead (broken) adult and two immature specimens. SnaiL Cave. Sec. 4, T3N, RI4E, MDM, USGS Columbia Quad.; Stan- ford Grotto, 73, Ray Bennett coll., March 10, 1951. ‘Taken in twilight zone.” This eave and a number of others discovered have been given names for the first time. Snails were the only form of life found in Snail Cave. Monadenia mormonum (Pfeiffer). One living and one dead adult. CRYSTAL STANISLAUS Cave. Sec. 23, T3N, RI4E, MDM, USGS Colum- bia Quad.; Stanford Grotto, 56, Edward Danehy coll., November 25, 1950. Snails found “30 ft. in from entrance (twilight zone).” Also Western Spel. Inst., C-64, C-166, and C—169, Raymond deSaussure, eoll., January 30, 1954. Found in various locations from the mouth of the cave to the twilight zone inside. Monadenia mormonum (Pfeiffer). Two living immature specimens and 10 dead adults, some of large size, plus five immature shells, also dead. GRAPEVINE GuLCH Cave. Sec. 22, T3N, R14E, MDM, USGS Columbia Quad.; Stanford Grotto, 95, Edward Danehy and Jack Jacobus colls., Vout. XXIX] A. G. SMITH: SNAILS FROM CALIFORNIA CAVES 4] iG: at 18. 19! April 17, 1951. “Found in twilight zone.” Also, Western Spel. Inst., C-3, C-5, and C-—7, Raymond deSaussure, coll., February 2, 1953. Monadena mormonum (Pfeiffer). A total of 21 adult shells and six immature specimens from three collectings, all dead. Also one living juvenile specimen. Helminthoglypta cypreophila (Binney & Bland). Two adults and three juvenile dead specimens. BUCKEYE CAVE. One of the Grapevine Gulch caves; Western Spel. Inst., C-6, Raymond deSaussure coll., February 2, 1953. From the twilight zone. Monadenia mormonum (Pfeiffer). Six dead specimens. Woot Hoituow Cave. Another of the Grapevine Gulch group. Ray- mond deSaussure coll., May 30, 1954, Western Spel. Inst. C—212, C216, and C—217. Taken in several locations from the entrance to the zone of total darkness. Monadenia mormonum (Pfeiffer). One immature living snail and eight other dead shells, five of which are adults. One of the last, broken and heavily inerusted, was found deposited with subaerial cave coral (PI. 1, fig. 1). KENNY’s GROTTO. Grapevine Gulch group of caves. Western Spel. Inst., C-215, Raymond deSaussure coll., May 20, 1954. From the floor in the twilight zone. Monadema mormonum (Pfeiffer). Five adult and four imma- ture shells, all dead. Helminthoglypta cypreophila (Binney & Bland). Two dead im- mature specimens. Music Hatt Cave. Between Moaning Cave and Parrott’s Ferry; Western Spel. Inst., C-2, Raymond deSaussure coll., February 1, 1953. Taken in total darkness at the foot of a 150-foot fissure. Monadenia mormonum (Pfeiffer). Fifteen adult and three imma- ture specimens, several almost completely covered or imbedded in calcite crystal masses, a number of others clean or mud covered. SPELEOGEN Cave. Near Music Hall Cave; Western Spel. Inst., C-—4, C-32, and C—50, Raymond deSaussure ecoll., February 2, April 4, and Aprils, 1953: Monadenia mormonum (Pfeiffer). Six adults and one immature, all dead, one found 160 feet down in a fissure. Helminthoglypta cypreophila (Binney & Bland). A single living juvenile taken on the floor in the twilight zone. 42 bo w bho coda | CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH Ser. TUOLUMNE CouUNTY Rating Cave. See. 34, T3N, R14E, MDM, USGS Columbia Quad.; Western Spel. Inst., C-67, C—77, C-81, and C-—82, Raymond deSaus- sure coll., May 16, 1953. Taken in various locations in the eave. Monadenia mormonum (Pfeiffer). Three living immature snails, plus eight adult and four juvenile dead specimens. Helminthoglypta cypreophila (Binney & Bland). A single dead juvenile. Pine Loe Cave. See. 36, T3N, R14E, MDM, USGS Columbia Quad.; Western Spel. Inst., C-87, Raymond deSaussure ecoll., July 18, 1953. Snail shells collected throughout the cave in the zones of twilight to total darkness. Monadenia mormonum (Pfeiffer). Eight adult and four juvenile shells, dead. Haplotrema alameda Pilsbry. One broken adult. SMALL Cave. Sec. 36, T3N, R14E, MDM, USGS Columbia Quad.; Western Spel. Inst., C-88, Raymond deSaussure ecoll., July 13, 1953. From the twilight zone in the front part of the cave. Monadenia mormonum (Pfeiffer). One adult and one immature, dead. Helminthoglypta cypreophila (Binney & Bland). One dead ju- venile. Haplotrema alameda Pilsbry. Three adults and two juveniles, dead. Goutp- ToorH Caves. See. 14, T3N, R14E, MDM, USGS Columbia Quad.; Western Spel. Inst., C-161, Raymond deSaussure eoll., Janu- ary 30, 1954. On the ground, under rocks, daylight to twilight zones. Monadenia mormonum (Pfeiffer). One broken adult and eight im- mature dead specimens. Helminthoglypta cypreophila (Binney & Bland). One dead ju- venile. CRYSTAL TUOLUMNE Cave. Sec. 21, TIN, RI6E, MDM, USGS Sonora Quad.; Western Spel. Inst., C-183, Raymond deSaussure coll., Febru- ary 6, 1954. On the ground just outside the entrance. Monadenia tuolumneana Berry. Four adult and one immature, all dead specimens. (C.A.S. 34678.) Helminthoglypta cypreophila (Binney & Bland). One dead adult. Vou. XXIX] A. G. SMITH: SNAILS FROM CALIFORNIA CAVES 43 30. 31. 32. CrystaL ButterRFLY Cave. Sec. 21, TIN, RI6E, MDM, USGS Sonora Quad.; Western Spel. Inst., C-184, Raymond deSaussure coll., Feb- ruary 6, 1954. Just inside the entrance in the twilight zone. Monadenia tuolumneana Berry. One dead adult. (C.A.S. 34679.) CRYSTAL PALACE Cave. Sec. 26, T3N, R14E, MDM, USGS Columbia Quad.; Western Spel. Inst., C-60 and C—64, Raymond deSaussure and Bill Wise ecolls., April 17 and May 24, 1953. In twilight zone. Monadenia mormonum (Pfeiffer). One living immature snail and four dead (two adult) examples. Cave Man Cave. Sec. 28, TIN, RI6E, MDM, USGS Tuolumne Quad.; Western Spel. Inst., C-380, Raymond deSaussure coll., April 2, 1953. Monadenia circumcarinata (Stearns). : \, : A ay iE ws . | NSE 162 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. stood by a knowledge of the past distribution of a similar forest which was composed of many of the same woody genera. Fossil records indi- eate that a deciduous mesophytie forest extending as far north as Man- ehuria, Greenland, and Alaska existed during the early and middle parts of the Tertiary period over a considerable portion of the northern hemi- sphere. This forest belongs to the so-called Arcto-Tertiary flora. The northwestern part of North America has provided a fairly com- plete chronological record of the history of the Areto-Tertiary flora in this region. The changes in the composition of this flora have been out- lined by Chaney (1940) on the basis of a series of fossil floras. The vari- ous fossil floras from Eocene through Miocene times in western North America show that there were for any period of time latitudinal and alti- tudinal variations in composition of the vegetation, but that as the climate changed from warm and wet in the earlier part of the Tertiary to cool and dry in the later part, the vegetation changed also. The climatic trend toward lower temperature and precipitation resulted in the elimi- nation of an evergreen neotropical flora, which existed as far north as Oregon, by the end of the Oligocene, and in the southward migration of a temperate deciduous flora of higher northern latitudes to replace it. By early Miocene times the redwood and broad-leaved, deciduous trees were dominant in Oregon. The Bridge Creek flora of the John Day Basin, late Oligocene or early Miocene, contains (Chaney, 1944) redwoods and broad-leaved deciduous trees characteristic of a wet summer climate. Other Miocene floras of the same age were of similar composition. How- ever, during the Miocene, the Cascades were uplifted and with the changes in topography and climate, the redwoods and broad-leaved trees began to disappear. The Maseal flora of later Miocene showed that trees such as black and live oaks, box elders, madrono, syeamores, and poplars began to replace the trees requiring more moisture. By the Pliocene, redwoods and other moisture requiring trees had nearly disappeared. Hydrangea, which belonged to this mesophytie element, is not reported after the Miocene in western North America. LaMotte (1952) lists six fossil species of Hydrangea from Eocene, Oligocene, and Miocene deposits in southern Alaska, Washington, Oregon, California, Idaho, and Colorado. These species and their places of col- lection are as follows: H. alaskana Hollick. Paleocene or Eocene. Jaw Mountain, Alaska Pen- insula, Alaska. H. californica MacGinitie. Eocene. Ione Gravels: Chalk Bluffs, Nevada. County, California. H. russellii Chaney & Sanborn. Eocene-Oligocene. Fisher: Goshen, Lane County, Oregon. H. fraxinifolia (Lesquereux) Brown. Oligocene. Florissant, Colorado. VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 163 H. reticulata MacGinitie. Oligocene. Weaverville, Hay Fork, Trinity County, California. H. bendiret (Ward) Knowlton. Miocene. Maseall: Grant County, Ore- gon. Latah: Spokane, Washington. Whitebird, Idaho. Trout Creek, Harney County, Oregon. Hydrangea is thus seen to be a member of several fossil floras in west- ern North America from early to middle Tertiary (see Map 1). Its typi- cal occurrence was in the temperate mesophytic forest characteristic of the Arcto-Tertiary flora; however, it is known also from one of the warm humid neotropical floras present in the Pacific Northwest during early Tertiary times. Its presence in the Goshen flora indicates that the sub- tropical element of the genus, occurring now in Central and South Amer- ica, may have had a northward extension in the Pacific Northwest. As this neotropical flora died out and was replaced by temperate floras, the subtropical element of Hydrangea was also lost in the area. These records of Hydrangea in the Tertiary floras of the west consti- tute the only fossil records of this genus in North America. In eastern North America, Tertiary fossil plants are known only from the south- eastern Gulf Coast. Although Hydrangea occurs today in the eastern United States, it is not found in any of these deposits. In China, Hydrangea is reported from the Miocene Shanwang flora in Shantung Province by Hu and Chaney (1940), who show that the ma- jority of Shanwang genera are wide ranging in China today and that this fossil flora includes elements which now characterize the forests of eentral and south China and the Yangtze Valley. Many of these are temperate northern genera, and although they are found today in middle and upper levels in the mountains, they are rare or absent from lower altitudes in these areas. The Shanwang flora is the only record of Mio- cene vegetation in China. As in western North America, Hydrangea is not found in China in deposits later than the Miocene. Pliocene floras of the northern interior of China lack the mesophytic elements present in the Shanwang flora. Hu and Chaney also showed that the modern forests of Korea and Japan were very similar to the Miocene vegetation of northern China, but that Pliocene floras from Japan are not different from central Japanese forests of today. THE SPECIES IN CULTIVATION Hydrangeas were cultivated in China and Japan long before their introduction into Europe and America, and many of our garden forms eame from the Orient. There is no way of knowing how long these forms had been grown by the Chinese and Japanese; we merely know that many Europeans, including Thunberg, Siebold, Fortune, and Maximowicz, found 164 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. these plants as well established horticultural subjects in China and Japan before they sent them to Europe. As with other horticultural plants hav- ing forms which are the result of many generations of selection, their origin is lost in obscurity. A discussion of the origin of the cultivated forms of Hydrangea macrophylla is given under that species. Haworth-Booth (1950) has discussed the horticultural aspects of the genus in his recent book, The Hydrangeas. Since Haworth-Booth has covered the subject adequately, the present discussion of Hydrangea as a garden plant will only list the species cultivated, together with some of the synonyms by which they are often known. Complete synonymy is given under each of the species mentioned. 1. Hydrangea macrophylla. Of the species listed, this one has always been the most popular because of its versatility, both as a garden plant and as a florists’ plant. Hydrangea macrophylla subsp. macrophylla var. macrophylla (H. Hortensia, H. hortensis, H. opuloides, Hortensia opu- loides). This is the form to which belong the many globose-headed plants with all sterile flowers called hortensias. Hydrangea macrophylla subsp. macrophylla var. normalis is the form to which belong the plants with a few sterile flowers surrounding a cluster of fertile flowers. Haworth-Booth calls these lacecaps. Hydrangea macrophylla subsp. serrata (H. serrata) is said to be in cultivation, but it is so similar to subsp. macrophylla that it is very difficult to recognize garden forms as actually belonging to this subspecies. 2. Hydrangea arborescens subsp. arborescens. A form of this sub- species with all sterile flowers, called H. arborescens grandiflora, is often cultivated. Subspp. discolor (H. cinerea) and radiata (H. radiata) are oe- easionally cultivated. 3. Hydrangea quercifolia. 4. Hydrangea paniculata. A form of this species with all sterile flowers, known as H. paniculata grandiflora, is called the ‘“‘peegee” hydrangea. 5. Hydrangea heteromalla (H. xanthoneura, H. Bretschneideri, H. vestita). 6. Hydrangea anomala subsp. petiolaris (H. petiolaris, H. scandens). This plant is often erroneously called Schizophragma hydrangeoides, a quite different Asiatic climber. 7. Hydrangea involucrata. A form of this species with inflorescences consisting of all sterile flowers has been cultivated in Japan. It appar- ently has not been introduced into the United States. 8. Hydrangea aspera subsp. aspera (H. villosa). H. aspera subsp. strigosa (H. strigosa, H. aspera var. macrophylla). H. aspera subsp. ro- busta (H. longipes). H. aspera subsp. Sargentiana (H. Sargentiana). 9. Hydrangea serratifolia (H. integerrima). The only species of the section Cornidia to be cultivated. VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 165 SYSTEMATIC TREATMENT Hydrangea Linnaeus, Genera Plantarum 189. 1754. Hortensia Commerson ex Jussieu, Genera Plantarum, 214. 1789. Cornidia Ruiz & Pavon, Florae Peruvianae et Chilensis Prodromus, 53, t. 35. 1794. Sarcostyles Presl ex Seringe in DeCandolle, Prodromus, 4:15. 1830. Much branched shrubs, erect, or scandent and clinging by means of aerial rootlets; deciduous or evergreen; inflorescences, young branchlets, and leaves variously pubescent; leaves opposite, simple, chartaceous or leathery, penninerved, margins entire, denticulate, or serrate, rarely pin- nately lobed; inflorescences usually terminal, occasionally axillary, on current year’s wood, consisting of compound, much branched clusters, rounded or pyramidal in shape, made up of complicated, dendritically branched systems, the ultimate members of which bear a single flower or small clusters of two to seven flowers; flowers fertile or sterile; fertile flowers small, perfect, usually very numerous, comprising the main body of the inflorescence; calyx tubes adnate to ovary, forming the hypan- thium, apices flat or conical; sepals 4 or 5, inconspicuous, deltoid; petals 4 or 5, ovate or spatulate, valvate, deciduous during anthesis, falling separately or rarely together as a cap; stamens 8 or 10, rarely as many as 20, filaments usually as long as petals or longer, filiform, anthers oblong to somewhat rounded; ovary inferior or half superior, 2 to 4, rarely 5 locular, styles 2 to 4, rarely 5, separate when ovary is inferior, united at their bases on top of conical apical disk of the hypanthium when ovary is half superior, stigmas terminal to more or less decurrent on inner surface of styles; ovules numerous, borne on axile placentas; cap- sules hemispherical or turbinate, dehiscent at apex between the styles; seeds numerous, small, linear or elliptic, sometimes caudate, rarely cir- cular with a wing surrounding body of seed; sterile flowers consisting of 3 to 5 enlarged petaloid sepals, usually lacking other flower parts, conspicuous, few in number, arranged on the periphery of the inflores- cences, or in some cultivated forms numerous and comprising the entire inflorescence. The generic name, Hydrangea, is derived from two Greek roots, hydor, meaning water, and angeion, diminutive of angos, meaning vessel or con- tainer, in allusion to the shape of the mature capsules. GEOGRAPHICAL DISTRIBUTION: eastern Asia, from the Himalayas east- ward to central and southern China, Sakhalin, Japan, the Ryukyu Islands, Formosa, the Philippine Islands, Sumatra, and Java; eastern North America; mountains of Central America southward into the Andes of South America. TYPE SPECIES: Hydrangea arborescens Linnaeus. 166 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. KEY TO THE SECTIONS, SUBSECTIONS, AND SPECIES la. Plants deciduous, erect or scandent, unopened inflorescence not enveloped by broadly ovate floral bracts (except in H. involucrata). Section Hydrangea 2a. Ovary inferior in both anthesis and fruit, capsules truncate at apex, styles separate to their bases. 3a. Seeds with rounded wing surrounding body of seed, petals falling to- gether as a cap or calyptra; plants scandent, clinging by means of aerial rootlets. Subsection Calyptranthe 6. H. anomala 3b. Seeds without rounded wing, petals falling separately, plants erect. 4a. Seeds caudate. (Plants of China and Japan.) Subsection Asperae 5a. Leaves pinnately lobed. 3. H. sikokiana 5b. Leaves not lobed. 6a. Bracts broadly ovate to rounded, few, enveloping the un- opened inflorescence, leaving noticeable scars on falling; capsules pubescent. 4. H. involucrata 6b. Bracts lanceolate, numerous, scattered through the inflo- rescence, covering but not enveloping the unopened in- florescence; capsules glabrous. 5. H. aspera 4b. Seeds not caudate. (Plants of the southeastern United States.) Subsection Americanae 7a. Inflorescence a flat-topped cluster; leaves not lobed. 1. H. arborescens 7b. Inflorescence a pyramidal cluster; leaves pinnately lobed. 2. H. quercifolia 2b. Ovary half superior in both anthesis and fruit; capsules conical at their apices. 8a. Seeds caudate; petals truncate at their bases; styles in fruit shorter than conical disk of capsule. Subsection Heteromallae 9a. Inflorescence a pyramidal cluster, calyx lobes 0.5-1 mm. long. 9. H. paniculata 9b. Inflorescence a flat-topped cluster, calyx lobes 0.8-3 mm. long. 10. H. heteromalla 8b. Seeds not caudate; styles in fruit longer than conical disk of capsule; petals clawed at their bases. Subsection Petalanthe 10a. Inflorescences consisting entirely of fertile flowers, stamens 2 to 3 times as long as styles, leaves with serrations 3-8 mm. deep. 7. H. hirta 10b. Inflorescences consisting of both fertile and sterile flowers; sta- mens same length as styles to as much as two times their length, leaves entire or with serrations less than 3 mm. deep. 8. H. scandens 2c. Ovary inferior in anthesis, becoming half superior in fruit; capsules conical at their apices. Subsection Macrophyllae 11. H. macrophylla 1b. Plants evergreen, shrubs or lianes clinging by aerial rootlets; leaves leathery; floral bracts broadly ovate, few, enveloping the unopened inflorescence, leaving Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 167 noticeable scars on falling. (Plants of Mexico, Central and South America, and one species in the Philippine Islands and Formosa. ) Section Cornidia lla. Inflorescence consisting of a single, terminal, flat-topped or rounded cluster; flowers pink or white; steriie flowers present or absent. Subsection Monosegia 12a. Calyx lobes and petals 4 or 5. 13a. Leaves oval, broadest near middle; sterile flowers present; calyx lobes broadly deltoid, less than 0.5 mm. long. 14a. Branchlets of inflorescence glabrous. (Plants of the Sierra Madre Occidental in the state of Durango, Mexico.) 12. H. Seemannii 14b. Branchlets of inflorescence pubescent with stellate hairs borne on short epidermal papillae. (Plants of Philippine Islands and Formosa.) 14. H. integrifolia 13b. Leaves obovate, broadest above middle; sterile flowers not present; calyx lobes deltoid to semi-orbicular, 1-2 mm. long. (Plants of Guatemala. ) 19. H. Steyermarkii 12b. Calyx lobes and petals 4. (Plants of Central and South America, from Costa Rica southward into the Andes.) 15a. Flowers white; leaves generally 5-14 cm. long, 3—5.5 em. wide; branchlets and inflorescence and lower leaf surface ferruginous tomentose with stellate hairs, which may be partly deciduous with age. 13. H. asterolasia 15b. Flowers pink; leaves generally 10-25 em. long, 4-15 em. wide; branchlets and inflorescence pubescent with single or stellate hairs, hairs 0.1—0.5 mm. long. 16a. Styles usually 3; sterile flowers absent. 18. H. Preslii e®16b. Styles usually 2; sterile flowers usually present. 17a. Styles during anthesis 0.3-1 mm. long and without evident stigmatic papillae; stamens 2—4.5 mm. long. 15. H. Oerstedii 17b. Styles during anthesis 1-2 mm. long and with evi- dent stigmatic papillae; stamens 0.4-3 mm. long. 18a. Stamens 0.4-1.6 mm. long, shorter than or rarely as long as styles; styles always separate. 16. H. peruviana 18b. Stamens 1.5-3 mm. long, longer than styles; styles usually coherent in bud and early anthesis, sepa- rating later. 17. H. diplostemona 11b. Inflorescences consisting of a series of rounded clusters, one above another; flowers white; sterile flowers always absent. (Plants of the Andes of South America. ) Subsection Polysegia 19a. Styles coherent in bud and occasionally during anthesis. 21. H. tarapotensis 19b. Styles always separate. 20a. Stamens 1-1.5 mm. long, not more than one and one-half times as long as styles. 22. H. Jelskii 20b. Stamens 3-7 mm. long, three to six times as long as styles. 21a. Inflorescence and lower leaf surface densely ferruginous pubescent with stellate hairs. (Plants of northern Peru.) 23. H. Mathewsii 168 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. 21b. Inflorescence and lower leaf surface glabrous or lightly pubescent with stellate or simple hairs. (Plants of central Chile and adjacent Argentina. ) 20. H. serratifolia Section I. HYDRANGEA Hydrangea, sect. Huhydrangea Maximowicz, Mémoires, Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16):6 (Revisio Hydrangearum Asiae Orientalis), 1867. Hydrangea, subgen. Hortensiella Koehne, Deutsche Dendrologie, 187, 1893. In part. Hydrangea, subgen. Euhydrangea (Maximowicz) Schneider, Handbuch der Laub- holzkunde, 1:384, 1905. Deciduous shrubs, usually erect, rarely scandent and clinging by aerial rootlets; leaves membranaceous or chartaceous; bracts lanceolate, scattered throughout inflorescence (except in Hydrangea involucrata), bracts ocea- sionally absent; stamens usually 10; ovary inferior or half superior; capsule truncate or conical at apex; seeds elliptic to oval, sometimes caudate. TYPE SPECIES: Hydrangea arborescens Linnaeus. The repetition of the generic name for the section containing the type species is in accordance with a rule adopted by the Nomenclature See- tion of the International Botanical Congress held in Paris in 1954 (Stafleu, 1954). Subsection 1. AMERICANAE Hydrangea, sect. Huhydrangea, subsect. Americanae (Maximowicz) Engler, in Engler & Prantl, Natiirlichen Pflanzenfamilien, 3, 2a:204, 1930. As Group, Rehder, Manual Cultivated Trees and Shrubs, p. 291, 1927; as Series, Rehder, Manual Cultivated Trees and Shrubs, ed. 2, p. 286, 1940. Hydrangea, sect. 1, Euhydrangea, ser. 1. Petalanthe, Americanae Maximowicz, Mémoires Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16): 6 (Revisio Hydrangearum Asiae Orientalis), 1867. Flowers white; ovary inferior; capsules truncate at apex; seeds not caudate. North American species. TYPE SPECIES: Hydrangea arborescens Linnaeus. 1. Hydrangea arborescens Linnaeus, Species Plantarum, 937. 1753. Much-branched shrub, 1 to 3 m. tall; deciduous; young branchlets and inflorescence strigose pubescent, or puberulent; leaves ovate, elliptic- ovate, or broadly ovate, acute or acuminate at apex, rounded, subcordate, or cuneate at base, broadest near center or toward base, margins serrate, 6-16 em. long, 2.5-12 em. wide, their length usually 1.5 to 2.5 times their width, lower surface glabrous or nearly so, or pilose or tomentose, upper surface glabrous, or with occasional scattered hairs; inflorescence a rounded, compound, much-branched cluster, 5-10 em. wide; fertile flowers white, very numerous; hypanthium 0.8-1 mm. long; calyx lobes 5, deltoid, VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 169 0.5 mm. long; petals 5, oblong, 2 mm. long, falling separately; stamens 10, 1.5-4 mm. long, the 5 longer ones about twice as long as the shorter ones; ovary inferior; styles 2, or rarely 3, 1 mm. long; capsule 2—-2.5 mm. long, 2-3 mm. wide at apex, somewhat conical at apex, styles 1.2-1.5 mm. long, ribs of capsule 8 or 9, conspicuous; seeds elliptical, blunt at ends, 0.5—0.8 mm. long, 0.4-0.6 mm. wide, prominently striate; sterile flowers white, sometimes absent, usually 4-lobed, occasionally 3-lobed, lobes ovate or orbicular, 8-15 mm. in diameter. (See plate 4, figure 1.) DIsTRIBUTION: Eastern United States from southern New York south- ward to the southern Appalachian Mountains, westward to the Ozarkian Plateaus. (See map 2.) Key TO THE SUBSPECIES OF HyDRANGEA ARBORESCENS 1. Lower leaf surface green, glabrous or nearly so, with hairs, when present, only along the midrib and secondary veins; sterile flowers usually not present. a. subsp. arborescens 1. Lower leaf surface white or gray, variously pubescent; sterile flowers present or absent. 2. Lower leaf surface gray, velutinous, pilose, or tomentose, hairs usually not dense enough to be matted; hairs when examined under magnification of about 50 diameters are seen to be conspicuously roughened by numerous tubercles. b. subsp. discolor 2. Lower leaf surface usually white or silvery, rarely gray, with felt-like indumen- tum, hairs densely matted; hairs when examined under magnification of about 50 diameters are seen to be smooth or with only a few scattered tubercles. c. subsp. radiata la. Hydrangea arborescens Linnaeus subsp. arborescens. Hydrangea arborescens Linnaeus, Species Plantarum, 937. 1753. Hydrangea vulgaris Michaux, Flora Boreali-Americana, 1:268. 1803. A specimen under this name, without locality or other data, in the Michaux Herbarium of the Natural History Museum, Paris, may be the type. Hydrangea arborescens var. vulgaris (Michaux) Seringe in DeCandolle, Prodromus, 4:14. 1830. Hydrangea arborescens var. kanawhana Millspaugh, Preliminary Catalogue of the Flora of West Virginia, 363. 1892. Type from West Virginia, Millspaugh 291 (WVA). Hydrangea arborescens var. australis Harbison, American Midland Naturalist, 11:255. 1928. Type from Union Springs, Alabama, Harbison June 18, 1902. (Biltmore Herb. No. 1339i, US.) (Note: A number of additional synonyms, listed by Rehder in the Bibliography of Cultivated Trees and Shrubs, pp. 199-200 (1949) under Hydrangea arborescens, are not included here. These could not be verified; in some cases, the names were based on cultivated plants and the type specimens were not designated; in other cases, the type specimens could not be located.) Lower leaf surface glabrous or nearly so, with hairs, when present, along midrib; individual hairs conspicuously tubercular, 0.3-1 mm. long, 10-25 microns wide; sterile flowers often not present. 170 CALIFORNIA ACADEMY OF SCIENCES | Proc. 4TH SER. TYPE LOCALITY: Virginia. Type collection: ‘““Dr. Clayton, ex Ameriea, An. 1734 num. 79” (BM); ILLUSTRATIONS : The Botanical Magazine, vol. 13, table 437. 1799. Gleason, H. A. Illustrated Flora of the Northeastern United States and Adjacent Canada, 2: 275. 1952. New Flora and Silva, 4: figure 79, 1932 (habit); 9: figure 34, 1937 (habit), both as Hydrangea arborescens grandiflora. DISTRIBUTION: Southern New York southward through the Appa- lachian Mountains to the coastal plain of central Alabama, westward to the Ozarkian Plateaus. ALABAMA: Bullock, Cullman, DeKalb, Etowah, Hale, Henry, Lee, Madison, Montgomery, St. Clair, Tuscaloosa, and Win- ston counties. ARKANSAS: Baxter, Benton, Clay, Carroll, Craighead, Craw- ford, Faulkner, Independence, Logan, Marion, Perry, Pike, Polk, Pope, Saline, Washington, and Yell counties. Disrricr of COLUMBIA. GEORGIA: Bartow, Catoosa, Clarke, Clay, Cobb, DeKalb, Fannin, Floyd, Fulton, Gilmer, Gordon, Gwinnet, Habersham, Hall, Hancock, Lumpkin, Madison, Pickens, Rabun, Rockdale, Stephens, Towns, and White counties. ILLINOIS: Champaign, Jackson, Johnson, Madison, Randolph, Richland, Shelby, and Union counties. InpIana: Cass, Clinton, Crawford, Floyd, Harrison, Howard, Huntington, Jefferson, Knox, Lawrence, Marion, Martin, Mon- roe, Morgan, Ohio, Orange, Owen, Parke, Perry, Putnam, Ripley, Scott, Vermillion, and Warren counties. Kentucky: Allen, Anderson, Bath, Bell, Boyd, Boyle, Breathitt, Carter, Casey, Christian, Daviess, Edmon- son, Estill, Fayette, Greenup, Grayson, Hancock, Harlan, Jefferson, Jessamine, Johnson, Letcher, Lyon, Madison, McCreary, Ohio, Pike, Union, Warren, and Wolfe counties. LourstANa: West Feliciana Parish. Maryann: Allegany, Cecil, Frederick, Harford, Howard, Montgomery, and Prince Georges counties. Mississtppt: Tishomingo and Wayne counties. Missouri: Barry, Butler, Callaway, Christian, Crawford, Dent, Dunklin, Franklin, Gasconade, Greene, Hickory, Howell, Iron, Jefferson, MeDon- ald, Miller, Newton, Oregon, Ozark, Pulaski, Ripley, St. Charles, St. Genevieve, Shannon, Stoddard, Stone, Taney, Wayne, Webster, and Wright counties. New Jersey: Burlington, Mercer, and Warren counties. New York: Chemung, Steuben, and Tioga counties. NORTH CAROLINA: Alexander, Alleghany, Ashe, Avery, Buncombe, Cherokee, Craven, Cum- berland, Durham, Forsyth, Graham, Haywood, Jackson, Iredell, Macon, Madison, Montgomery, Orange, Polk, Rockingham, Stanley, Stokes, Surry, Swain, Transylvania, Wake, Watauga, Wilkes, Yadkin, and Yancey coun- ties. Onto: Adams, Belmont, Butler, Columbiana, Fairfield, Green, Ham- ilton, Jefferson, Lawrence, Meigs, Pike, Ross, and Scioto counties. OKLA- HoMA: Cherokee, Delaware, Johnston, Leflore, and McCurtain counties. Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA i q(al PENNSYLVANIA: Allegheny, Blair, Bradford, Bucks, Butler, Cameron, Centre, Chester, Clearfield, Elk, Fayette, Franklin, Fulton, Laneaster, Lehigh, Luzerne, Lycoming, Mifflin, Monroe, Montgomery, Northampton, Perry, Philadelphia, Schuylkill, Snyder, Sullivan, and Westmorland coun- ties. SoUTH CAROLINA: Oconee County. TENNESSEE: Bedford, Blount, Campbell, Carter, Cheatham, Cocke, Coffee, Cumberland, Davidson, Fen- tress, Franklin, Grainger, Hamilton, Knox, Pickett, Rhea, Sevier, Shelby, Sullivan, Unicoi, and Washington counties. Vircinta: Arlington, Au- gusta, Bedford, Botetourt, Carroll, Fairfax, Fauquier, Giles, Green, Greens- ville, Isle of Wight, James City, Loudoun, Louisa, Madison, Mecklenburg, Montgomery, Middlesex, Page, Patrick, Rappahannock, Roanoke, Rock- bridge, Rockingham, Scott, Shenandoah, Smyth, Southampton, Surry and York counties. West VirGINIA: Barbour, Boone, Braxton, Cabell, Cal- houn, Doddridge, Fayette, Grant, Greenbrier, Hampshire, Harrison, Jef- ferson, Kanawha, Marion, Marshall, Mercer, Mineral, Monongalia, Mon- roe, Morgan, Nicholas, Ohio, Pendleton, Preston, Raleigh, Randolph, Ritchie, Summers, Taylor, Tucker, Tyler, Upshur, Wayne, Webster, Wetzel, and Wood counties. lb. Hydrangea arborescens subsp. discolor (Seringe) MeClintock, Journal Arnold Arboretum, 37: 373. 1956}. Hydrangea arborescens var. discolor Seringe in DeCandolle, Prodromus, 4:14. 1830. Hydrangea arborescens var. Deamii St. John, Rhodora, 23:208. 1921. Type from Jackson County, Indiana, Deam 28122 (isotype A). Hydrangea arborescens var. Deamii f. acarpa St. John, 1. c. Type from Montier, Shannon County, Missouri. Bush 6116 (holotype GH, isotypes MO, US). Hydrangea Ashei Harbison, American Midland Naturalist, 11:256. 1928. Type ma- terial said to be from Montgomery County, Arkansas, collected by Ashe, but no actual specimen was cited. Placed here because the leaves were described as being pubescent. Hydrangea cinerea Small, Bulletin Torrey Botanical Club, 25:148. 1898. In part. Small cited five syntypes. Of these Lamson-Scribner, June 29, 1890 (NY, US), and Kearney, July 16, 1894 (US, NY, MO), both collected at White Cliff Springs, Tennessee, and Small, August 7-12, 1895 (A, BH), Ringgold, Catoosa County, Georgia, are subsp. discolor. Lower leaf surface velutinous, pilose, or tomentose, usually grayish; individual hairs conspicuously tubercular, 0.83-1 mm. long, 10-25 microns wide; sterile ray flowers present or absent. Type collection of Hydrangea arborescens var. discolor: A cultivated specimen collected by Seringe in 1824 (G). 1 Herbarium specimens belonging to this taxon were annotated before the writer had seen the specimen typifying Seringe’s varietal epithet. Because there was some doubt as to the proper appli- cation of this name previous to this time a combination based on St. John’s var. Deamii was used to annotate the specimens studied earlier. 172 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. DISTRIBUTION: Occurring with subsp. arborescens except in the north- eastern part of the range of that subspecies. ALABAMA: Bibb, Calhoun, Clay, De Kalb, Jackson, Jefferson, Lauderdale, Madison, Marshall, and Tus- ealoosa counties. ARKANSAS: Baxter, Benton, Carroll, Cleburne, Crawford, Cross, Garland, Hot Springs, Newton, Pulaski, Phillips, Poinsett, Pope, St. Francis, Stone, Van Buren and White counties. GEORGIA: Catoosa, Cobb, Dade, Walker and Whitfield counties. ILLINoIs: Calhoun, Cham- paign, Jackson, Jersey, Johnson, Peoria, Pope, St. Clair, Vermillion, and Williamson counties. INpIANA: Bartholomew, Brown, Gibson, Jackson, Knox, Morgan, Monroe, Perry, Posey, and Tippecanoe counties. KEN- TucKY: Livingston and Union counties. Missourr: Cape Girardeau, Car- ter, Dallas, Jefferson, Madison, Newton, Oregon, Ozark, Phelps, Pike, Reyn- olds, Ripley, St. Genevieve, St. Louis, Scott, Shannon, Stoddard and Wash- ington counties. OKLAHOMA: Leflore County. TENNESSEE: Blount, Carroll, Davidson, Decatur, Franklin, Grundy, Hamilton, Marion, Monroe, Obion, Polk, Roane and Van Buren counties. le. Hydrangea arborescens subsp. radiata (Walter) McClintock, Jour- nal Arnold Arboretum, 37: 373. 1956. Hydrangea radiata Walter, Flora Caroliniana, 251. 1788. Hydrangea nivea Michaux, Flora Boreali-Americana, 1:268. 1803. A specimen under this name, without locality or other data, in the Michaux Herbarium, Natural History Museum, Paris, is doubtless the type. Hydrangea nivea Michaux var. conformis Seringe in DeCandolle, Prodromus, 4:14. 1830. Hydrangea cinerea Small, Bulletin Torrey Botanical Club, 25:148. 1898. In part. Small cited five syntypes. Of these Curtiss 833 (June, 1882) (MO, US, GH, UC, BM, W, G, K), Chilhowee Mountains, Blount County, Tennessee, and Kearney, June 24, 1893 (NY), Chilhowee Gap, Blount County, Tennessee, are H. arbo- rescens subsp. radiata. Lower leaf surface covered with a conspicuous white felt-like indu- mentum, occasionally a grayish tomentum; individual hairs either smooth (without any tubercules), 1-3 mm. long, 5-10 microns wide, or with a few scattered tubercules, 1-3 mm. long, 5-20 microns wide; sterile flowers usually present. Type Locauity of Hydrangea radiata: “Carolina.” Type collection: Walter s.n. (BM). DISTRIBUTION: Occurring with the other two subspecies in the southern section of the Blue Ridge Mountains, extending to the piedmont area in the vicinity of the Seneea River in the northwestern corner of South Carolina. Plants with indumentum composed of smooth hairs: GEorGIA: Rabun County. NortH Carotina: Buncombe, Graham, Henderson, Jackson, VoL. XXIX|] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 173 Macon, Madison, Polk, Swain, and Transylvania counties. SoutH Caro- LINA: Anderson, Greenville, Oconee, and Pickens counties. TENNESSEE: Blount County. Plants with indumentum composed of inconspicuously tubercular hairs: NortH CarouiIna: Buncombe, Henderson, Macon, Polk, Swain, and Tran- sylvania counties. TENNESSEE: Blount County. VARIATION IN Hydrangea arborescens Variation in Hydrangea arborescens involves the kind and amount of vesture on the lower leaf surface, the presence or absence of sterile per- ipheral flowers, and the size and shape of the leaves. The flower parts show no variation. The vesture characters of the lower leaf surface have been used to divide the species into subspecies as follows: 1. Lower leaf surface glabrous or nearly so, with only a few scattered hairs along the midrib (plate 1, figure 1). Subsp. arborescens. 2. Lower leaf surface pubescent with erect or spreading, straight, rather firm hairs which vary in amount from sparsely seattered to dense enough to cover the surface and form a gray tomentum (plate 1, figures 2 and 3; plate 2, figure 3). The hairs are covered with numerous tuber- cules, giving them a roughened appearance (plate 3, figure 4), and are 0.5-1 mm. long and 10-25 microns wide. These hairs are found in subsp. discolor. The few hairs along the midveins of subsp. arborescens have also these same dimensions and markings. 3. Lower leaf surface with intertwined fine woolly hairs, usually very dense and forming a white, or rarely gray, felt-like indumentum (plate 1, figure 4). The hairs have a smooth surface (plate 3, figure 1) and are 1-3 mm. long and 5-10 microns wide. Subsp. radiata. 4. Lower leaf surface with indumentum resembling that of 3 above, except for its gray color (plate 2, figure 2) and with hairs which differ in having a few scattered tubercules (plate 3, figures 2 and 3). The hairs are 1-3 mm. long and 5—20 microns wide. These hairs appear more like the hairs of subsp. radiata and therefore the individuals having them are included in this subspecies. The geographical distribution of the three subspecies is given on Map 2. It is seen from this map that subsp. arborescens is the most widespread of the three; that subsp. discolor occurs in the western and southeastern, but not in the northeastern part of the range of subsp. arborescens; and that subsp. radiata occurs only in the southeastern part of ranges of the other subspecies, which is the southern Appalachian Mountains. The dis- tribution of the three subspecies in the southern Appalachians is shown on Map 3. In this area occur the two forms of subsp. radiata, the one having smooth hairs and the other having hairs with seattered tubercules. 174 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Here also occur individuals of subsp. discolor having both “sparse” and “dense” pubescence, as well as subsp. arborescens. Sterile flowers are present in the inflorescences of some individuals and absent in others. There is evidence to show from herbarium specimens that their presence is associated in part with type of pubescence and in part with geographic range. Of 231 individuals of subsp. arborescens in the eastern part of its range, from Virginia to South Carolina, 65, or 27 per cent, had sterile flowers, while in the western part of its range, from Tennessee and Kentucky to Missouri and Arkansas, of 243 indi- viduals, 167, or 69 per cent, had sterile flowers. In the western part of the range of subsp. discolor (Missouri, Arkansas, Indiana, Illinois, and Ohio), of 87 individuals, 14, or 17 per cent had sterile flowers, but in the eastern part of its range, Georgia, Alabama, and Tennessee, where the pubescence is “sparse” to “dense,” of 39 individuals, 36, or 92 per cent, had sterile flowers. In subsp. radiata, 50 individuals of the form having smooth hairs were examined, and 49, nearly 100 per cent, had sterile flowers, and of the form having a few tubereules on its hairs, 9 out of 19, or 47 per cent, had sterile flowers. Since all of the subspecies are found in the southern Appalachian Mountains, and often more than one have been found by different collectors at, or near, the same locality, it seemed of interest to determine whether more than one subspecies occur together, and if so, whether the form of subsp. radiata with seattered tubercules on its hairs would be found. In the summer of 1948, the writer visited this region and collected Hydrangea material in about 50 localities. It was found that for the most part the plants were uniform at a given locality. In only two localities were there more than one subspecies. In Madison County, North Carolina, about 414 miles west of Walnut Gap, there were typical plants of both subsp. arborescens and subsp. radiata. The other locality was in the Pisgah Na- tional Forest, near White Pine Annex Campground, Transylvania County, North Carolina. The three subspecies were in this locality along with several atypical individuals. No. 48-59B (see plate 2, figure 2) was like the one previously referred to subsp. radiata. There were several other individuals (Nos. 48-61B, 48-59A, see plate 2, figure 1, and No. 48-60, see plate 1, figure 5) in which the hairs were short (0.5-1 mm.) and had Map 2. Distribution in the eastern United States of oO Hydrangea quercifolia Hydrangea arborescens @ subsp. arborescens A subsp. discolor W@ subsp. radiata Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 175 176 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. numerous tubercules as in subsp. discolor, but their width was less (5-20 microns) than is typical for this subspecies. Another individual, No. 48-61C, placed in subsp. arborescens because of being nearly glabrous, had a few scattered hairs along the veins which were like those of the individuals of subsp. discolor just described. A search for similar atypical individuals in other localities in the Pisgah Forest area was not successful. It would appear that such indi- viduals are rare. Dunean (Castanea 15: 153, 1950) reported finding subspp. arborescens and radiata growing together in Rabun County, Geor- gia, but he found no “intermediates” in a large number of plants in- spected. Perhaps if the localities where those individuals of subsp. radiata having ineonspicuously tubereulate hairs have been found were searched, other intermediates would turn up. However, it would appear that intermediates are not common and those few which have appeared have not caused the character of the populations to change. 2. Hydrangea quercifolia Bartram, Travels, xviii, nom., 382, t. 6, deser., 408, nom. 1791; ed. 2, xii, 380, t. 6, 406, 1792.2 Species Plantarum, ed. 4, C. L. Willdenow, 2: 634. 17993. Hydrangea radiata Smith, Icones Pictae Plantarum rariorum, t. 12. 1791. Non Walter, 1788. Stout erect shrub, much branched from base, 1-2 m. tall; young branchlets and petioles densely tomentose with long, fine, reddish hairs; branches of inflorescence pilose with short, coarse hairs; leaves suborbicular or oval in outline, 8-25 em. long, mostly 5-lobed, the lobes coarsely serrate, petioles one-eighth to one-third as long as blades; upper leaf surface gla- 2 According to Merrill (Bartonia, 23:10-35, 1945), the citation of H. quercifolia given in Index Kewensis isnot correct. The citation here is Merrill’s correction. 3 Because of the doubt which is attached to Bartram’s Travels as a legitimate place to begin nomen- clature, Willdenow’s publication may be considered to be the earliest valid one for this name. See Merrill, 7. c. pp. 17 and 27. Map 3. Distribution in the southern Appalachian Mountains of Hydrangea arborescens @ subsp. arborescens subsp. discolor A Sparsely pubescent @ tomentose subsp. radiata H@ indumentum white and hairs smooth * indumentum gray and hairs inconspicuously tubercular McOCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA Vou. XXIX] | “Tallule Y uospt, % ow, 9A0D sapen Nese ulmunoW i 178 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. brous; lower leaf surface pilose or tomentose with two kinds of hairs— short, coarse, tubercular ones, and longer, fine, smooth ones which form an intertwining network; inflorescence a compound, much branched cluster, pyramidal in shape, 15-25 em. long; fertile flowers white, very numer- ous; hypanthium 1-1.5 mm. long; calyx lobes 5, deltoid, 0.5-0.8 mm. long; petals oblong, acute, concave, 2—-2.5 mm. long; stamens 10, 3-6 mm. long; styles 2, 3, or 4, 1-1.4 mm. long, somewhat thickened toward apex; cap- sules 1.5-2.5 mm. long, 2-2.5 mm. wide at apex, apex truncate, ribs 7 or 8, inconspicuous; styles 1.2-1.5 mm. long; seeds elliptical, 0.7-0.8 mm. long, 0.3-0.4 mm. wide, prominently striate, blunt at each end; sterile flowers present, usually a single one terminating each lateral branch of the inflorescence, four-lobed, lobes orbicular, 10-15 mm. in diameter. (See plate 4, figure 2.) TYPE LOCALITY: GrorGIA: Crawford County, Sweet Water Brook —= Knoxville Branch (see Merrill, Bartonia 23:27, 1945). TyPE COLLECTION: W. Bartram s. n. (holotype BM). ILLUSTRATIONS : Curtis’s Botanical Magazine, table 975, 1806. New Flora & Silva 4: fig. 80. 1932 (habit). Gleason, H. A. 1952. Illustrated Flora of the Northeastern United States and Adjacent Canada, 2: 275. DISTRIBUTION: Southeastern United States, on the piedmont and coastal plain of western Georgia, Alabama, and Mississippi. (See Map 2.) AvAaBAMA: Blount, Calhoun, Clarke, Coneeuh, Cullman, Dallas, DeKalb, Etowah, Franklin, Jackson, Jefferson, Marshall, Shelby, St. Clair, Tallapoosa, Tuscaloosa, and Winston counties. FLoripA: Calhoun, Gadsden, Jackson PLATE 1. Vesture types in Hydrangea arborescens. Figure 1. Subsp. arborescens. Lower leaf surface glabrous except for few scat- tered hairs along veins. (from McClintock 48-44) Figure 2. Subsp. discolor. Lower leaf surface sparsely pubescent. (from Pal- mer 12640) Figure 3. Subsp. discolor. Lower leaf surface with slightly more hairs than in figure 2. (from Deam 2248) Figure 4. Subsp. radiata. Lower leaf surface covered with white felt-like indu- mentum. (from McClintock 48-53) Figure 5. Subsp. discolor. Lower leaf surface moderately pubescent with hairs finer than those in figure 3. (from McClintock 48-60) The photographs in this plate and in Plate 2 were taken by Richard Johnson. Magnification in all figures on both plates is 19x. Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 179 180 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. and Liberty counties. Grorcta: Bibb, Chattahooche, Clarke, Clay, Craw- ford, Dade, Decatur, Early, Floyd, Harris, Heard, Merriwether, Polk, Randolph, Stewart, Talbot, Taylor, and Upson counties. Louistana: East Feliciana and West Feliciana parishes. Mississippi: Adams, Attala, Copia, Grenada, Hinds, Holmes, Itawamba, Tallahatchie, Tishomingo, and War- ren counties. TENNESSEE: Lincoln, Marion, and Shelby counties. Discussion: Hydrangea quercifolia is remarkably uniform throughout its range. It is characterized by its pyramidal inflorescence and its lobed leaves, by which characters it is distinguished from H. arborescens. In geographic distribution these two species differ, although they overlap in part in the southern part of the range of H. arborescens. Hydrangea quercifolia is found chiefly in the area of the piedmont and coastal plain of western Georgia, Alabama, and Mississippi. Hydrangea arborescens subsp. arborescens and subsp. discolor occur infrequently in this area. Subsection 2. ASPERAE Hydrangea, sect. Euhydrangea Maximowicz, subsect. Asperae Rehder, Plantae Wil- sonianae, 1:39. 1911. Hydrangea, sect. Euhydrangea Maximowicz, subsect. Piptopetalae Maximowicz, Mémoires Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16):8 (Revisio Hydrangearum Asiae Orientalis), 1867. In part. Hydrangea, subgenus Euhydrangea, sect. Japonico-sinensis, subsect. Piptopetalae Schneider, Handbuch der Laubholzkunde, 388, 1905. In part. Flowers white or blue; ovary inferior; capsule truncate at apex; styles free to their bases; seeds caudate. Chinese and Japanese species. Type species: Hydrangea aspera Don. PLATE 2. Vesture types in Hydrangea arborescens and Hydrangea aspera. Figure 1. Hydrangea arborescens subsp. discolor. Lower leaf surface moderately pubescent with hairs somewhat finer, longer and more dense than those on Plate 1, figure 5. (from McClintock 48-59A) Figure 2. Hydrangea arborescens subsp. radiata. Lower leaf surface covered with gray indumentum, hairs not as dense as in Plate 1, figure 4, and having scat- tered tubercules. (from McClintock 48—-59B) Figure 3. Hydrangea arborescens subsp. discolor. Lower leaf surface gray to- mentose. (from McClintock 48-82.) Figure 4. Hydrangea aspera subsp. strigosa. Hairs straight, appressed along raised veins. (from Fang 15568) Figure 5. Hydrangea aspera subsp. aspera. Hairs tending to be erect and spreading. (from Fang 19721) 181 MONOGRAPH OF GENUS HYDRANGEA McCLINTOCK: Vou. XXIX] 182 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. 3. Hydrangea sikokiana Maximowicz, Bulletin scientifique Académie im- périale des Sciences de St. Pétersbourg, 31:42. 1887. Shrub, 1-2 m. tall; branchlets, inflorescence and pedicels pubescent with upwardly appressed hairs 0.5-1 mm. long; leaves pinnately lobed with 4 to 6 lobes, 8-21 em. long, 8-20 em. wide, lower surface pubescent along veins and veinlets with hairs 1-1.5 mm. long, upper surface with seat- tered hairs along veins; petioles 2-18 em. long, pubescent; inflorescence a flat-topped, compound, several-branched cluster, 12-30 em. across; bracts few, scattered through inflorescence, lanceolate to ovate, 10-30 mm. long, pubescent as leaves; fertile flowers purplish; hypanthium 1—1.6 mm. long; ealyx lobes 5, broadly deltoid, 0.5 mm. long; petals 5, 2-4 mm. long, 1.5— 2 mm. wide, truncate at base; stamens 10, rarely 8 or 9, 3.5-5.5 mm. long; styles 2, 1-1.5 mm. long; ovary inferior; capsule 2-3 mm. long, styles 1.5-2 mm. long; seeds elliptical, 1 mm. long, tapering at each end into minute tail-like extensions; sterile flowers always present, purplish, 4-lobed, lobes rounded, 1-3 em. in diameter. (See plate 5, figure 1.) TYPE LOCALITY: Japan, Honshu Island, Kii Province (now Wakayama and Mie prefectures). Type collection: Tanaka. Specimen not seen by the writer. DISTRIBUTION: Woods and mixed forests on Honshu and Kyushu islands in Japan, from 100 to 1,000 meters. (See Map 4.) JAPAN. HonsHu IsuaNp: Nara Prefecture: Tonomine, Mt. Odaigahara. Wakayama Pre- fecture: Mt. Koya. Kyusuu Isuanp: Oita Prefecture: Bungo. Kagoshima Prefecture: Mt. Hirishima. Kumamoto Prefecture: Hitoyoschichi. Discussion: Hydrangea sikokiana is distinguished from H. mvo- lucrata, the only other closely related Japanese species, by its pinnately lobed leaves. The American species, H. quercifolia, and H. sikokvana are the only species in the genus which have lobed leaves. PLATE 3. Surface appearance of hairs from lower leat surfaces of Hydrangea arborescens. Figure 1. Smooth hairs of subsp. radiata. (from McClintock 48-27) Figure 2. Hairs with very few tubercules from subsp. radiata. (from Cor- rell 3717) Figure 3. Hairs with few tubercules from subsp. radiata. (from Davis 7683) Figure 4. Hairs with numerous, conspicuous tubercules from subsp. discolor. (from Palmer 35469) Drawings made with a camera lucida. Magnification 350. neg Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA FIG. 1 oo % ‘. 2 ; ft z # \ aN } > & 183 184 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. 4. Hydrangea involucrata Siebold, Nova Acta: Academia (caesarea) Leopoldino-Carolina, 14(2):691. 1829. Hydrangea longifolia Hayata, Journal, College of Science, Imperial University, Tokyo, 25(19):91. 1908. Type from Formosa, Kawakami & Nakahara 690, not seen, photo CAS. Shrub, 1-2 m. tall; branchlets, inflorescence and pedicels pubescent with upwardly appressed hairs 0.5 mm. long; leaves ovate, 10-26 em. long, 5-17 em. wide, their length 1.5 to 2.5 times their width; both surfaces pubescent with appressed hairs; petiole 1.5-8 mm. long; inflorescence rounded, compound, many-branched; involucral bracts subtending the inflorescence, 1.5—2.5 em. long, broadly obovate, pubescent with upwardly appressed white hairs on outside, glabrous on inside, caducous, leaving a row of scars at base of inflorescence when falling; in addition, a few bracts, ovate or lanceolate in shape, are scattered through the inflorescence; fertile flowers lavender, hypanthium 1.2-1.5 mm. long; calyx lobes 5, deltoid, 0.2-0.6 mm. long; petals 5, 2-3 mm. long, truncate at base; stamens 10, rarely 11 or 12, 3.5-6 mm. long; ovary inferior; styles 2 or 3, 1-2 mm. long; capsules 3-4 mm. long, pubescent, styles 2-2.5 mm. long; seeds elliptical, 0.8-1 mm. long, tapering at each end into minute tail-like extensions; sterile flowers always present, lavender, 4-lobed, lobes rounded, 1-3 em. in diameter. (See plate 5, figure 5.) PLATE 4. Flowers and capsules of some American species of Hydrangea. Figure 1. Hydrangea arborescens. a. flower (without petals and some of sta- mens. Biltmore herb. 1339c). b. capsule (Biltmore herb. 8212d). Figure 2. Hydrangea quercifolia. a. flower. b. capsule. Figure 3. Hydrangea asterolasia. a. flower (Skutch 3342). b. capsule (A. Smith H470). Figure 4. Hydrangea peruviana. a. flower (Allen 534). b. capsule (Standley 39080). Figure 5. Hydrangea Mathewsii. Flower (Mathews 3051). Figure 6. Hydrangea Preslii. a. flower (Weberbauer 6776). b. flower (Skutch 2630). Figure 7. Hydrangea diplostemona. a. flower (Spruce 4328). b. young cap- sule (Spruce 4328). Figure 8. Hydrangea tarapotensis. a. flower (Smith 2627). b. young capsule (Lawrence 775). Figure 9. Hydrangea integrifolia. a. flower (Clemens June 1924). b. capsule (Wilson 9691). Figure 10. Hydrangea serratifolia. a. flower (Morton Dec. 3, 1906). b. capsule (Sargent Jan. 12, 1906). VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 185 186 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. TYPE LocALITty: Japan. Type collection: Collected in Japan by von Siebold. The specimens collected in Japan by von Siebold are in the Rijksherbarium in Leiden. Presumably the type of H. involucrata is there but the writer has not seen it. ILLUSTRATION: Siebold and Zucearini. Flora Japonica, plates 63 and 64. 1839 or 1840. DistRIBUTION: Woods and mixed forests on Honshu Island in Japan, from near sea level to 1,500 meters, and on Formosa. JAPAN. HoNsHuU IsLAND: Chiba Prefecture: Owari, Kiyozumi-yama. Fukushima Prefee- ture: Mt. Haguro. Gifu Prefecture: Norikura, Washiga-take. Gumma Pre- fecture: Ikaho. Kanagawa Prefecture: Yokohama, Miyanoshita, Hakone Mts., Kamakura, Mt. Takao. Nagano Prefecture: Usui-toge, Mt. Izuna, Mt. Tsubakura, Kuramoto, Kiso near On-take-san, Asamayama. Shiga Prefecture: Shinano. Tochigi Prefecture: Nikko. Tokyo Prefecture: Tokyo, Mt. Takao, Hachijo. Yamanashi Prefecture: Motsuko. (See Map 4.) Discussion: Hydrangea involucrata bears a close resemblance to H. aspera in flower and fruit structures, and in the shape of the leaves. The two species may be distinguished by the tomentose capsules and the in- volueral bracts of H. involucrata. These bracts which envelop the un- opened inflorescence resemble those found in the section Cornidia. The distribution of Hydrangea involucrata as listed above includes localities from Japan only. The species, however, occurs in Formosa as indicated by a specimen collected by Kawakami and Nakahara at Taito on the southeast coast in January 1906. This specimen, the type of Hydrangea longifolia Hayata, was not seen by the writer. A photograph of the specimen shows the sears left by deciduous involucral bracts, the presence of which identifies it with H. amvolucrata. PLATE 5. Flowers and capsules of some Asiatic species of Hydrangea. Figure 1. Hydrangea sikokiana. a. flower (Sakurai August 16, 1909). b. cap- sule (Komeda E53). c. seed (Komeda E53). Figure 2. Hydrangea heteromalla. a. flower (Wilson 2184). b. capsule (Li 23). c. seed (Li 23). Figure 3. Hydrangea paniculata. a. flower (Chung & Sun 469). b. single petal. c. capsule (Canton Christian College herb. no. 8738). d. seed (as preceding). Figure 4. Hydrangea anomala. a. flower (Wilson 4901). b. united petals or calyptra (as preceding). c. capsule (Steward et al. 502). d. seed (as preceding). Figure 5. Hydrangea involucrata. a. flower (Dorsett & Morse 980). b. capsule (Suzuki December 31, 1948). c. seed (Suzuki December 31, 1948). VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 187 y Vent ve ttn we tty Ret wea) OH 188 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. 5. Hydrangea aspera D. Don, Prodromus Florae Nepalensis, 211. 1825. Shrub or small tree, 1-4 m. tall; branchlets and inflorescence pubescent with stiff and upwardly appressed or curling and erect hairs; leaves variable in shape, lanceolate to ovate, broadly ovate or deltoid, the lanceo- late or ovate ones cuneate or rounded at base, 5-35 em. long, 1.5-15 em. wide, their length 2.5 to 7 times their width, serrate or serrulate along margin, petioles 0.5-6.5 em. long, the broadly ovate or deltoid ones rounded or cordate at base, 10-25 em. long, 7-17 em. wide, serrate or doubly ser- rate along margin, petioles 4-21 em. long; vesture on lower leaf surface variable and similar to that of stem; inflorescence a flat-topped, many- branched cluster, 10-30 em. broad; bracts lanceolate, densely hairy on outer surface, glabrous on inner, caducous or remaining; fertile flowers white or blue, numerous, hypanthium 1—1.5 mm. long; sepals 5, 0.4-0.5 mm. long; petals 5, 2-2.5 mm. long, truncate at base, deciduous before stamens; stamens 10, 2-7 mm. long, usually alternately long and short; ovary com- pletely inferior; styles 2, 3 or 4, 0.5-1.5 mm. long; capsules 3-5 mm. long; seeds elliptical, 0.7-0.8 mm. long, 0.3-0.4 mm. wide, tapering at each end into minute tail-like extensions; sterile flowers always present, white, 4-lobed, lobes 1-3.5 em. long, usually slightly longer than wide, with entire or fimbriate margins. (See plate 6, figure 6.) DistriIBuTION: In woods or thickets, eastern Himalaya Mountains from 1,500 to 2,300 meters; western China from 700 to 4,000 meters; central and eastern China from 250 to 2,100 meters; Formosa, Sumatra and Java from 1,000 to 2,500 meters. (See Map 4.) KEY TO THE SUBSPECIES OF HYDRANGEA ASPERA 1. Stems, lower leaf surface, and petioles strigose or velutinous. 2. Leaves lanceolate or narrowly ovate, their length usually two to five times as great as their width, rounded at base; petioles wsually 1-5 cm. long. 3. Lower leaf surface strigose, hairs appressed, straight. b. subsp. strigosa Map 4. Distribution in eastern Asia of Hydrangea aspera HM subsp. strigosa A subsp. aspera (inset Sumatra and Java) Vv subsp. robusta * subsp. Sargentiana @ subspecies not determined A Hydrangea sikokiana o Hydrangea involucrata MONOGRAPH OF GENUS HYDRANGEA 189 McCLINTOCK: Vou. XXIX] 190 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47TH SER. 3. Lower leaf surface velutinous, hairs erect, somewhat curled, occasionally long coarse pili along midvein, petiole and stem. a. subsp. aspera 2. Leaves ovate or broadly ovate, usually less than two times as long as wide, base truncate or cordate; petioles wsually 5-17 cm. long. c. subsp. robusta 1. Stem, lower leaf surface along midveins, and petioles velutinous, and in addi- tion, interspersed with conspicuous, fleshy trichomes, 2—5 mm. long, split at apex; leaves large, 15 cm. or more long, lower surface velutinous. d. subsp. Sargentiana 5a. Hydrangea aspera D. Don subsp. aspera. Hydrangea aspera D. Don, Prodromus Florae Nepalensis, 211. 1825. Hydrangea oblongifolia Blume, Bijdragen tot de Flora van Nederlandsch Indié, 920. 1826. Type from Java, Blume, s. n. (isotypes W-U, NY). Hydrangea vestita var. fimbriata Wallich, A Numerical List of Dried Specimens of Plants in the East India Company’s Museum, published in 1828. Name only. Type from Nepal, Wallich in 1821 (K—W, listed in Wallich’s Catalogue under No. 440). Hydrangea longipes Hemsley var. lanceolata Hemsley, Journal Linnaean Society Botany, 23:274. 1887. Type from Patung, Hupeh, Henry 1786. (K, according to Rehder in Plantae Wilsonianae 1:40, this is the type.) Hydrangea aspera var. strigosior Diels, Botanische Jahrbiticher, 29:375. 1900. Type from Szechuan, Bock & von Rosthorn 2546 (holotype O). Hydrangea Kawakamii Hayata, Journal, College of Science, Imperial University, Tokyo, 25(19):90. 1908. Type from Formosa, Kawakami & Mori 1875 (type not seen, photo CAS). Hydrangea aspera var. scabra Rehder, Plantae Wilsonianae, 1:31. 1911. Type from Kangting (Tatsienlu), Sikang, Wilson 4485 (holotype A). Hydrangea aspera var. velutina Rehder, 1. c. 30. Type from Wa-Shan, Szechuan, Wilson 2405 (holotype A, isotypes K, E). Hydrangea fulvescens Rehder, 1. c. 39. Type from Wushan, Szechuan, Wilson 1373 (holotype A, isotypes EK, K, W). PLATE 6. Flowers and capsules of some Asiatic species of Hydrangea. Figure 1. Hydrangea hirta. a. flower (Wilson 6958). b. petal (as preceding). c. capsule (Wilson 7680). d. seed (as preceding). Figure 2. Hydrangea macrophylla subsp. serrata. a. flower (Henry M83). b. petal (as preceding). c. immature capsule (as preceding). d. mature capsule (Wilson 7621). e. seed (as preceding). Figure 3. Hydrangea macrophylla subsp. macrophylla, immature capsule (Su- zuki July 1917). Figure 4. Hydrangea scandens subsp. chinensis. a. flower (Tso 21020). b. petal (as preceding). c. capsule (Steward & Cheo 1003). Figure 5. Hydrangea scandens subsp. liukiuensis. a. flower (Beattie & Kuri- hara 10646). b. capsule (Wilson 7797). c. seed (as preceding). Figure 6. Hydrangea aspera. a. flower (Wilson 2401). b. capsule (Wilson 4337). Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 1s 192 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Hydrangea glabripes Rehder, I. c. 30. Type from Fang Hsien, Hupeh, Wilson 2391 (holotype A, isotypes E, K, W). Hydrangea villosa Rehder, 1. c. 29. Type from Wenchuan Hsien, Szechuan, Wilson 1227 (holotype A, isotypes E, K). Hydrangea Rehderiana C. K. Schneider, Handbuch der Laubholzkunde, 2:940. 1912. Type from Nanto, Hupeh, Wilson 1533 (holotype W, isotypes NY, K). Premna Merinoi Léveillé, Sertum Yunnanense, 3. 1916. Type from “brousses des montagnes a Siai-Ho, 2700 m.”, Yunnan, Maire May-July, 1912 (fragment of holotype A). Leaves lanceolate or narrowly ovate, their length usually two to five times as great as their width, rounded at base; petioles usually 1-5 em. long; lower leaf surface velutinous, hairs erect, or somewhat curled, oec- easionally long coarse pili along midvein, petiole and stem. (See plate 2, figure 5.) Type LocaLity of Hydrangea aspera: Narainhetty, Nepal. Type col- lection: Buchanan September 26, 1802 (holotype BM). ILLUSTRATIONS : H. H. Hu & W. Y. Chun, 1927. Icones Plantarum Sinicarum, fase. I, plate 28. New Flora and Silva 7: plate opposite p. 177. 1935, as H. villosa. B. Hayata, 1908. Flora Montana Formosae in Journal, College of Sei- ence, Imperial University, Tokyo 25(19), plate 8, as H. Kawakami Hayata. DISTRIBUTION: CHINA. ANHweI: Kimen, Wang Shan. CHEKIANG: T’ien-mu Shan, King Yuan, Suian, Yentang Shan, Taishun. Hunan: Anfu, Sinning, Changning, Sikwangshan district. HupeH: Changyang, Ch’ing young Hsien, Enshih Hsien, Fang Hsien, Hsing Shan Hsien, He-tan-kon, Ichang, Liang sung you, Peh Yang Tsai, Mo Yu Ping, Nanto, Paokang, Patung, Sui Yeh See, Siao-Ya-tze, Wushan, Wuchang. Krtanost: Kiu Chi Hsien, Yungshiu. Kwanest: Pai-yun-an, Ling-chwan district. KWANGTUNG: Lokchong district. KwrtcHow: Kweiting, Pi-tsieh-Hsien, Fan Ching Shan, Tsunyi Hsien, Tungtze. Sr1kaNnG: Lushan via Paohing (Mouping), Paohing, Yaan. SZECHUAN: Chung Hsien, Hsi-Chang between Huangnipu and Yaan, Huangnipu, Kwanhsien, Lungan, Nanchuan, Omei Shan, Opien, Wanhsien, Yun-ching Hsien. YUNNAN: Salwin Valley near Champutong, Io chan, Kiao me ti, Ku-long-tschang, Lanping, Lungling, Lu-se, Lichiang, Mengtsz, Po- shang, Pingpien, Shuang-chiang, west flank Shweli-Salwin divide, banks of Shweli River, Shunning, Shang pa, between Tengyueh and Lungling, Weisi, Wenshan. BURMA. Bhamo district near Sinlun Kaba, Kachin Hills. BHU- TAN. Tunbje Punakha. INDIA. West BENGAL: Darjeeling. SIKKIM: Rumbhi Jhora. TIBET. Tsarong, Ngawchang Valley. SUMATRA. Habin- saran, Seriboe, road from Siantar to Brastagi. FORMOSA. Arisan, Kiraiki. Vout. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 193 5b. Hydrangea aspera D. Don subsp. strigosa (Rehder) McClintock, Journal Arnold Arboretum, 37: 373. 1956. Hydrangea strigosa Rehder, Plantae Wilsonianae, 1:31. 1911. Hydrangea aspera D. Don var. macrophylla Hemsley, Journal Linnaean Society, Botany, 23:273. 1887. Type from ‘“Nanto and mountains to the northward,” Hupeh, Henry s. n. (K). Hydrangea aspera D. Don var. angustifolia Hemsley, 1. c. 273. Type from Ichang, Hupeh, Henry s. n. (K). Hydrangea aspera D. Don var. typica Diels, Botanische Jahrbiicher, 29:375. 1900. Type from Nanchuan, Szechuan, Bock & von Rosthorn 299 (holotype O). Hydrangea aspera D. Don var. sinica Diels, l. c. 375. Type from Nanchuan, Szechuan, Bock & Rosthorn 629 (holotype O). Hydrangea Hemsleyana Diels var. lanceolata (Hemsley) Diels, J. c. 376. (Based on H. longipes Hemsley var. lanceolata Hemsley. ) Hydrangea aspera D. Don var. strigosior Diels, l. c. Type from Wei kuan, Szechuan, Bock & von Rosthorn 2546 (holotype O). Hydrangea villosa Rehder var. Mairei Léveillé, Catalogue des Plantes du Yunnan, 254. 1917. Type from Kiao me ti, Yunnan, Maire July—October 1912 (isotype A). Hydrangea villosa Rehder forma sterilis Rehder, Journal Arnold Arborteum, 11:161. 1930. Type from Changyang, Hupeh, Wilson 1473a (holotype A, isotypes K, EH, Wale Leaves lanceolate or narrowly ovate, their length usually two to five times as great as their width, rounded at base; petioles usually 1—5 em. long; lower leaf surface strigose, hairs appressed, straight. (See plate 2, figure 4.) Type Locatity of Hydrangea strigosa: Techang, Hupeh, China. Type collection: Wilson 765 (holotype A, isotypes K, US, E, W). ILLUSTRATIONS : New Flora and Silva, 9: figure 33, 1937 (habit), as Hydrangea strigosa var. macrophylla. Curtis’s Botanical Magazine, table 9324, 1933, as H. strigosa Rehder. DistripuTION: CHINA. Huprn: Changyang, Chienshih, Ichang, On tan tsian. Kwanost: Tzu Yuen district. KwricHow: Kweiting, Lao Shan, Luneli, Pichieh, Pin fa, Pi-tsieh-Hsien, Tating, Touchan. Srkana: Kang- ting, Lushan via Paohing (Mouping), Ta Kwan. SzecHuan: Chia-ting, Kwanhsien, Omei Shan, Opien, Tien-Tsian, Wa-Shan, Wenchwan Hsien, Weiknan. YuNNAN: Champutong, Chih-tse-lo, Chungtien on bank of Yangtze, Ku-long-tschang, Liang-shan, Lichiang, Long ky, Lon pon, Lou Lan, Ma chou, Menetsz, Paoshan, Pingpien, Sichour-hsien, Shang pa, Tao Shan between Chao and Hungai, Tengyueh, Tchen fong chan, Tsekou, Tsan Shan Range near head of Yangpi Valley, Tunghai, Weisi, between Yangpi and Yungchang-fu, mountains of Yangpi River drainage basin, Yaoan. 194 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. BURMA. Hpimaw. NEPAL. Barpak Village, Narainhetty, Thanjey. JAVA. West side of Jang Plateau at Besoeki, Dieng Plateau, Mt. Kendeng, Mt. Malabar, Preanger, Mt. Soembing, Tjibodas on Mt. Gedah. FOR- MOSA. Mt. Taihei, Mt. Nanko, Nai-taroko. 5e. Hydrangea aspera Don subsp. robusta (Hooker f. & Thomson) Me- Clintock, Journal Arnold Arboretum, 37: 373. 1956. Hydrangea robusta Hooker f. & Thomson, Journal Linnaean Society of London, Botany, 2:76. 1858. Hydrangea robusta var. Griffithii C. B. Clarke in J. D. Hooker, Flora of British India, 2:404. 1878. Type from “Eastern Bhotan,” Griffith 2495 (holotype K, isotype P). Hydrangea longipes Franchet, Nouvelles Archives, Muséum d’histoire naturelle, Paris, ser. 2, 8:228 (Plantae Davidianae, 2:45). 1885. Type from Paohing (Moupin), Sikang, David in 1869 (holotype P, isotypes P). Hydrangea longipes, Hemsley, Journal Linnaean Society of London, Botany, 23:273. 1887. Type from Hupeh, Henry s. n. (K). Hydrangea Rosthornii Diels in Engler’s Botanische Jahrbiicher ftir Systematik, 29:375. 1901. Type from Nanchuan, Szechuan, Bock & von Rosthorn 471 (isotype W). Hydrangea Hemsleyana Diels, 1. c. 376, based on Hydrangea longipes Hemsley, not Franchet. Hydrangea aspera var. longipes (Franchet) Diels, 1. c. 374, based on Hydrangea longipes Franchet. Hydrangea Mavrimowiczii Léveillé, Bulletin Académie international de Géographie Botanique, 12:114. 1903. Syntypes from Kweichow, Martin & Bodinier 1654 (H, P), and Cavalerie 22 (FE). Leaves ovate or broadly ovate, usually less than two times as long as wide, base truncate or cordate; petioles usually 5-17 em. long; lower leaf surface strigose, hairs appressed, straight. Type LocaLity of Hydrangea robusta: Sikkim. Type collection: J. D. Hooker s.n. (holotype K, isotypes K, NY, BM, C, G, P, W, 8S). ILLUSTRATIONS : H. H. Hu and W. Y. Chun, Icones Plantarum Sinicarum, fase. III, plate 140, as Hydrangea longipes Rehder, plate 141, as Hydrangea Ros- thorn Diels. DistripuTION: CHINA. Honan: Lushih, Sunghsien. Hunan: Anfu. Hureu: Changlo Hsien, Changyang, Chienshih, Hsing Shan Hsien, Hiang sung you, Ichang, Nanto, Paokang, Patung, Wan Tsao Chan. Kansv: Lower Tebbu country between Hera Sambaku & Dayaya gorge, Wantsang. Kwanest: Chuen Yuen, Kweilin district, Sankiang Hsien. KWEICHOWw: Lao Shan, Tunyun. SuHanst: Locality not stated. SHenst: Mt. Kin-quan- san, Tai-pei-shan. Skane: Kangting, Paohing, Wu Ya Ling. SzecHUAN: VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 195 Hungya, Nanchuan, Omei Shan, Opien, Wa-shan, Wenchwan Hsien, Weik- nan. YUNNAN: Kiukiang Valley (Taron), Long ky, Shang pa, Tchen fong ehan. BURMA. Bhamo. BHUTAN. Mirichoma Timpu, Trashiyangsi. INDIA. Urtar PrapescH: Garhwal. West BenGAauL: Darjeeling. SIKKIM: Khursiang, Rhikisun, Sureil. TIBET. Yigrong Valley. FORMOSA. Ari- san, Ariko-Banti. 5d. Hydrangea aspera Don subsp. Sargentiana (Rehder) McClintock, Journal Arnold Arboretum, 37: 373. 1956. Hydrangea Sargentiana Rehder, Plantae Wilsonianae, 1:29. 1911. Stem, lower leaf surface along midveins, and petioles velutinous, and in addition, interspersed with conspicuous, fleshy trichomes, 2-5 mm. long, split at apex; leaves broadly ovate, 15 em. or more long. Type Locauity: Hsing Shan Hsien, Hupeh Province, China. Type dpa?) collection: Wilson 772 (holotype A). ILLUSTRATIONS: H. H. Hu & W. Y. Chun, 1933. Icones Plantarum Sinicarum, fase. ITI, plate 138. New Flora and Silva, 12: figure 39, 1940 (habit), and 9: figure 32, 1937 (habit). Curtis’s Botanical Magazine, table 8447. 1912. DistRIBUTION: CHINA. Hupen: Hsing Shan Hsien, Wan Tsao Chan. Discussion oF Hydrangea aspera Hydrangea aspera is a wide-ranging species, occurring from the eastern Himalayas across western and south-central China and on the islands of Formosa, Sumatra, and Java. It varies throughout this range in size and shape of the leaves and in the pubescence of the lower leaf surface. Most of the names listed in the synonymy were based on such variants. The leaves vary in shape from lanceolate to ovate, often broadly ovate; their bases are cuneate, rounded or truneate, or occasionally cordate. In actual size they vary in length from 8 em. to as much as 30 em. The petiole leneth varies also, and is often associated with leaf shape. Thus a broadly ovate leaf usually has a longer petiole than one which is lanceolate. The pubescence may be strigose with straight hairs which are appressed and arranged along the veins and veinlets more or less at right angles to them in a regular fashion, or arranged along the veins somewhat irregularly, or the pubescence may be velutinous with curled erect hairs. Although various combinations of leaf shape and vesture occur, most individuals fall into the following categories: 1, leaves lanceolate or ovate, and 2, 196 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. leaves ovate; often broadly so. In the first of these categories, the pubes- cence may be either strigose, referred here to subsp. strigosa, or velutinous, referred here to subsp. aspera. In the second category, the pubescence is mostly strigose and appressed, referred here to subsp. robusta. Most of the herbarium material examined can be referred to these three taxa, but there are some specimens which clearly are intermediate. A fourth group, subsp. Sargentiana, has large, broadly ovate leaves, but its chief distinguishing character is the presence of conspicuous, more or less turgid trichomes on the stem, petioles, and lower leaf surface along the midvein. Don stated in his deseription of Hydrangea aspera that this species was based on a collection made by “Hamilton” at Narainhetty, Nepal. In the British Museum (Natural History) there is a specimen from the Lam- bert Herbarium collected by “Dr. Buchanan” at Narainhetty, September 26, 1802. It may be considered to be the type of H. aspera. Stapf in discussing the variation of Hydrangea strigosa Rehder (Cur- tis’s Botanical Magazine, 156: tab. 9324, 1933) referred to the relation of this taxon to H. aspera Don, stating that the two may be given specific rank or be treated as minor groups within H. aspera sensu lato (1. H. aspera protonyma and H. aspera strigosa). Stapf clearly stated that he recognized the first alternative. Since Stapf did not accept the cate- gory referred to by the name H. aspera strigosa, his publication of this combination is not valid according to Article 43 of the International Rules which states that a name is not valid which is not accepted by the author who published it. Subsection 3. CALYPTRANTHAE Hydrangea, sect. Hydrangea, subsect. Calyptranthe (Maximowicz) McClintock, Journal Arnold Arborteum, 37:373. 1956. Hydrangea, sect. Calyptranthe Maximowicz, Mémoires Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16):6 (Revisio Hydrangearum Asiae Orientalis), 1867. Schneider, Handbuch der Laubholzkunde, 1:384, 1905, as subgenus. Flowers white; petals cohering at apex and falling together as a cap or calyptra; stamens 10, 15 or 20; ovary inferior; capsule truncate at apex; styles free to their bases; seeds with a rounded wing surrounding body of seed. Only one species, Hydrangea anomala D. Don, widespread in eastern Asia. 6. Hydrangea anomala D. Don, Prodromus Florae Nepalensis, 211. 1825. Woody climber; branchlets and inflorescence sparsely pubescent with upwardly spreading hairs, 0.5 mm. long; leaves ovate, cuneate, truncate, or cordate at base, acute or acuminate at apex, 6-15 em. long, 3.5-10.5 em. wide, their length 1.5 to 2 times their width, glabrous except for scattered Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 197 hairs or clumps of hairs along midvein on lower surface; inflorescence a flat-topped, several-branched cluster; involucral bracts lanceolate or ovate- lanceolate, membranous; fertile flowers white; hypanthium 1.5-2 mm. long; calyx lobes 5, 0.5-0.9 mm. long; petals 5, 2-3 mm. long, remaining at- tached and falling as a cap before the stamens elongate, cap apiculate or rounded at apex; stamens 9-20, 3-5 mm. long, 3 to 4 times as long as styles; ovary inferior; styles 2 or 3, 1-1.3 mm. long; capsule glabrous, 2.5-4.5 mm. long, styles 1-3 mm. long; seeds 1.5-2 mm. long, 1-1.5 mm. wide, nearly round, winged all around; sterile flowers usually present, white, sepals rounded, 1-2 em. in diameter. (See plate 5, figure 4.) DISTRIBUTION: Woods and mixed forests in Japan from sea level to 2,000 meters, on Honshu, Kyushu and Hokkaido islands; in Korea on Dagelet and Quelpaert Islands; in the eastern Himalaya Mountains of northern India, from 1,200 to 2,500 meters, and in west central to east. central China, from 600 to 3,800 meters. (See Map 5.) Kry TO THE SUBSPECIES IPS EA TIVE TSS OY UO milli. eet ree a oe a on ee ee eee cee subsp. anomala lee Stamens a5) to 20) 2 ee SF dre Oa ee a subsp. petiolaris 6a. Hydrangea anomala D. Don subsp. anomala. Hydrangea anomala D. Don, Prodromus Florae Nepalensis, 211. 1825. Hydrangea altissima Wallich, Tentamen Florae Napalensis, 2: t. 50. 1826. Type from Nepal, Wallich Catalogue No. 439/1. Stamens 9 to 15. ILLUSTRATIONS: N. Wallich, Tentamen Florae Napalensis, 2: t. 50, 1826. As Hydrangea altissima. H. H. Hu & W. Y. Chun, Icones Plantarum Sinicarum, fase. III, plate 142. 1933. DISTRIBUTION: Eastern Himalayas and western and central China. CHINA. ANuHwer: Hwang Shan. CuHEKkiane: Tihtai Shan; Kingyuan. Hunan: Yun Shan near Wukang. Hupen: Hsin Tien-tsze; Paokang; Iehang; Hingshan; Patung. Kwanest: Ling Wan district; Lucheng; Shan Chuen to Tzeyuan district. KwericHow: Fan Ching Shan; Shihtsien district. SzecHUAN: Omei Shan; Nanechwan; Kwanhsien; Wushan; O’Pien Hsien; Ma-pien Hsien; Wa-shan; Tchen-keou-tin. YUNNAN: Tali Range; Mt. Gitsa west of Mekong and north of Weisi; Yungshan; Weisi; Shang-pa; Kungshan (Chamutang, Champutong); Niangonang; Mingkwong Valley; Kani Pass, N. Maikha-Salween divide; Shweli-Salween divide; Ma Nung Mt. near Tali. INDIA. Assam: Talung Dzong, Balipara frontier. Urrar [ Proc. 4TH SER. CALIFORNIA ACADEMY OF SCIENCES *(SpUBIS] OUTAAI[IYd JOSUL) sysuauryo ‘dsqns suapunos nabuniphiy Me pue (SedAv[eullzZ Ut9}SeM JOST) DIJDWOUD DabUDAPAH @® JO vIsy Udosvo UL UOTINGIAISIG ‘“G dVIN Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 199 PrapvesH: Dehra Dunn Dist., Deohan; Chakrata; Siwalik and Jaunsar divi- sions; Chekerli, Siwahk and Jaunsar divisions; Landour; Dhanolti; Garh- wal Dist., Tehri Dist. Sikkim: without definite locality. HmacHaL Pra- DESH: Chamba Dist.; Simla district. Punsasp: Chenab Valley on ascent to Pass Punjab; Kangra Dist.; Nagkanda; Dharmsala; Kulu; Jibhi; Bushire; Manali in Lahul subdivision; Kanugadh; Kumaun dist. Wrest BENGAL: Darjeeling. NEPAL. Doti; Montes Shandaghery, Sherpore; Sermore; Thare. TIBET-BURMA FRONTIER. Seinghku Wang; Adung Valley; Kangfang route. BHUTAN. Tilageng Punakha; Kandam Punakha; Shamb- lng, Kuru chu valley; Transmiyangse Dzone. 6b. Hydrangea anomala D. Don subsp. petiolaris (Siebold & Zucearini) datas L956; ( Hydrangea petiolaris Siebold & Zuccarini, Flora Japonica, 1:108, t. 54. 1839 or 1840. McClintock, Journal Arnold Arboretum, : Hydrangea ? bracteata Siebold & Zuccarini, 1. c. 172, t. 92. 1841. No type specimen was designated for this nor the following name. The plates may be considered to represent the types in each case. Hydrangea cordifolia Siebold & Zuccarini, J. c. 113, t. 59, fig. 2. 1839 or 1840. Hydrangea scandens Maximowicz, Mémoires Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16):16. 1867. Non Seringe 1830, nec Poeppig 1830. Based on the above two taxa of Siebold and Zucecarini. Hydrangea glabra Hayata, Journal College of Science, Imperial University, Tokyo, 25 (art. 19):89, plate 6. 1908. Type from Formosa, Kawakami & Mori 1787. (Type not seen, photo CAS.) Hydrangea tilaefolia Léveillé, in Fedde’s Repertorium Novarum Specierum 8:282. 1910. Three syntypes, Taquet 809 (K, E, G), Fawrie 1654 (H, G, BM), 358 (B, BM), all from Dagelet Island, off the south coast of Korea. Stamens 15 to 20. TYPE LOCALITY: Japan. Type specimen: No specimen is mentioned. The plate (Flora Japonica 1: t. 54) may, therefore, be considered to rep- resent the type. ILLUSTRATIONS : B. Hayata. Flora Montana Formosae in Journal, College of Science, Im- perial University, Tokyo, 25 (art. 19): plate 6. 1908. As Hydrangea glabra Hayata. Curtis’s Botanical Magazine, plate 6788. 1884. As H. petiolaris. Siebold and Zucearini. 1839 or 1840. Flora Japonica, 1 : t. 59, figure IT, as H. cordifolia, and t. 54, as H. petiolaris. New Flora and Silva, 1: figure 43, 1929 (habit). New Flora and Silva, 9: figure 31, 1937 (habit). This, and the above illustration, as H. petiolaris. 200 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47TH SER. DISTRIBUTION: SOVIET FEDERATED SOCIALIST REPUBLIC. Sakhalin (Karafuto) : Oko. JAPAN. Honsuu Isuanp: Aomori Prefecture: Tanabu; Hakkoda; Mts. of Shichinoki. Osaka Prefecture: Mt. Kongosan. Akita Prefecture: Akita. Gifu Prefecture: Mt. Kasagi; Mitaka. Hyogo Prefecture: Tii-mura; Ashiu. Nagano Prefecture: Otake-gawa; Mt. Izuna; Shinano. Wakayama Prefecture: Sitigawa-mura. Yamagata Prefecture. Nara Prefecture: Tonamine; Mt. Misen. Kanagawa Prefecture: Mt. Ha- kone. Tochigi Prefecture: Chuzenji; Shiobara. Shizuoka Prefecture: Sho- jiko; Mt. Amagi on Idzu Peninsula. Iwate & Miyagi border, Mt. Kurikoma. Yamanashi Prefecture: between Fujiyoshida and Lake Shoji. Gumma Prefecture: Mt. Akagisan; Aze. HoxkKatpo Isuanp. Ishikari Subprefecture: Sapporo; Maru-yama. Abashiri Subprefecture : Kitami Fuji. Oshima Sub- prefecture: Hakodate. Hidacha Subprefecture: Erimo Cape. Iburi Sub- prefecture: Noboribetsu; Muroran. Shiribeshi Subprefecture: Moribetsu. Soya Subprefecture: Soya Cape (Cape Romanzoff). KyusHu ISLaANnp. Kago- shima Prefecture: Yaku-shima. Fukuoka Prefecture: Hikosan. SHIKOKU Istanp. Kochi Prefecture. Formosa. Arisan. Korea. Dagelet and Quel- paert islands. Discussion oF Hydrangea anomala Throughout its wide geographical range, Hydrangea anomala varies but little. The most obvious variation is in the number of stamens; the Himalayan and Chinese plants have 9 to 15, and the Japanese 15 to 20. The Japanese plants have been considered to be a separate species from the Himalayan and Chinese plants. However, the differences between the two groups do not seem to warrant specific status for them and they are here treated as subspecies. Don stated in the Prodromus that he based Hydrangea anomala on a collection made by Wallich in Nepal. Since there was no such specimen in the Lambert Herbarium in the British Museum it was assumed that it must be in the Wallich Herbarium at Kew. In the Wallich Herbarium there are three collections of H. anomala written up as follows under Wal- lich’s name for the species in his Numerical List: “439 Hydrangea altissima Wall. in Herb. 1824. 1. Napalia 1821 2. Kumaon, R. B. 3. Sermore, Lieut. Gerard.” Although most of the material in the Wallich Herbarium was brought by Wallich to England on his return from India in 1828 some had been sent ahead previous to this datet. Material under No. 439 was received 4 Wallich so stated this in his explanation of the purpose of his Numerical List; see Kew Bulletin for 1913, p. 256. Vout. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 201 in England in 1824 and was available to Don before work on the Prodromus was completed. A significant point in Don’s description of Hydrangea anomala is the statement that the flowers are small and uniform and that there are no sterile flowers. It may have been for this reason that Hooker and Thom- son in 1857° placed H. anomala in synonymy under Adamia cyanea, a species considered today to belong in the genus Dichroa. Don’s description of H. anomala further states that it has two short, thick, recurved styles. While it is true that Dichroa has no sterile flowers, it has more than two styles, and these are not short, thick, or recurved. Of the three specimens of ‘Hydrangea altissima” in the Wallich Herbari- um under No. 439, only one, No. 439/1, was collected by Wallich. This speci- men consists of a branch bearing two young inflorescences on which sterile flowers are not evident, and a detached portion of an inflorescence which is in anthesis and has sterile flowers. The label reads: ‘‘Sheopore. Montes Chandaghery. May 1821.” Since the specimen was collected by Wallich in Nepal, and consists of a branch without evident sterile flowers, it presumably was the specimen on which Don based his H. anomala. In the material recognized in this paper as H. anomala there were several specimens which lacked sterile flowers. The detached inflorescence men- tioned above may not have been part of the specimen which Don saw, but may have been added later when the specimen was mounted. Wallich published Hydrangea altissima in the Tentamen the year following the publication of Don’s Prodromus. H. altissima was illustrated on plate 50, but no description accompanies this plate and there is no indication of the specimen on which Wallich based his name. The plate of H. altissima appears to have been made from a painting in the East India Company’s collection in the herbarium at Kew. This, in turn, appears very similar to Wallich’s specimen No. 4389/1, with the exception of the fact that in the specimen, the flowers are mostly in bud, while the inflorescences shown on the painting have fully opened flowers. It seems conclusive that the specimen in the Wallich Herbarium, 439/1, served first as the basis of Don’s H. anomala, and later as the material for the painting in the East India Company’s collection from which the plate of H. altissima in Wal- lich’s Tentamen was adopted. Maximowiez in 1867 cited the three species of Siebold and Zucearini, listed here as synonyms of Hydrangea anomala subsp. petiolaris, under his name of H. scandens. In his discussion of these species of Siebold and Zucearini, Maximowiez mentioned particularly H. bracteata and the plate representing it in the Flora Japonica (vol. 1, table 92). This plate shows a branch, the leaves and bracts of which resemble those of H. anomala, but with flowers having 10 stamens and petals which are separated and 5 See Journal Linnaean Society Botany, 2:76. 202 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. not joined together to fall as a calyptra. Maximowicz pointed out that the Japanese plant of this relationship had 15 and not 10 stamens. He sug- gested that the illustrator had made a grave error in this illustration and that the error untortunately was accepted by the describers. This expla- nation may be as good as any for explaining the anomaly of the plant pictured. The plant is referred here questionably to H. anomala subsp. petiolaris because of its Japanese origin. Subsection 4, PETALANTHE Hydrangea, sect. Huhydrangea, subsect. Petalanthe (Maximowicz) Rehder, Plantae Wilsonianae, 1:37. 1911. Hydrangea, sect. 1, Euhydrangea, ser. 1, Petalanthe, Japonico-sinensis Maximowicz, Mémoires Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16):6 (Revisio Hydrangearum Asiae Orientalis). 1867. In part. Flowers white or blue; ovary half superior; capsule conical at apex; styles longer than conical disk; seeds not caudate. Chinese and Japanese species. TypE SPECIES: Hydrangea hirta Siebold & Zucearini. 7. Hydrangea hirta (Thunbere) Siebold, Flora, 11: 757. 1828. Viburnum hirtum Thunberg, Flora Japonica, 124. 1784. Shrub, 0.5—2 m. tall, branchlets and inflorescence pubescent with coarse, upwardly appressed hairs, 0.4-1 mm. long, and in addition, longer spread- ing hairs on branchlets; leaves 3-10.5 em. long, 2.5-7 em. wide, length equaling width to as much as 2.5 times as great, oval, deeply serrate, with serrations 3-8 mm. deep, both surfaces with scattered appressed hairs along veins and veinlets; petiole 1-3.5 em. long, pubescent with spreading hairs; inflorescence a rounded, compound cluster, 2-10 em. wide; bracts 1 or 2 pairs below inflorescence, 1-3 em. long, leaf-like, lanceolate, entire or with 2 to 4 deep serrations; fertile flowers only pres- ent, bluish; hypanthium 0.5-1 mm. long; calyx lobes 5, rarely 4, 0.5 mm. long; petals 5, rarely 4, 1.5-2 mm. long, reflexed before falling, cuneate at base; stamens 10, rarely 8, 2.5-4+ mm. lone, 2 to 3 times as long as styles; ovary half superior; styles 3 or 4, 1-1.5 mm. long, united for 0.3-0.5 mm., free for 0.6—-1 mm.; capsule 3-3.5 mm. long (including styles) ; hypanthium 1.5-2 mm. long, styles 1-2 mm. long, united for 0.5-1 mm., less than one- half of their length; seeds 0.7-0.9 mm. long, 0.4-0.5 mm. wide, elliptical, blunt at each end. (See plate 6, figure 1.) TYPE LOCALITY: Japan. Type collection: Thunberg (holotype UPS, photo A). ee) Vout. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 20 DISTRIBUTION : Mixed forests and woods, on Honshu Island in Japan, on Okinawa Island in the Ryukyu Islands, from 200 to 1000 meters. JAPAN. HonsHu Isuanp: Kyoto Prefecture: Atago County; Kyoto. Saitama Pre- fecture: Mitsumine; Chichibu; Mt. Takamidzu. Kanagawa Prefecture: Hakone Mts.; Hanno; Mt. Takao. Shizuoka Prefecture: Shiojiko (Fuji- yama region); Shimoda. Nagano Prefecture: Mt. Asama; Atago near Karuizawa; Gora to Miyanoshita; Kiso near Ontake-san. Tochigi Prefee- ture: Nikko; Chuzenji. Tokyo Prefecture: Mt. Takao: Mt. Sougakuzan. Gifu Prefecture: Hokunomura; Nakayama-hichiri. Hyogo Prefecture: Map 6. Distribution in Japan of @ Hydrangea hirta A Hydrangea scandens subsp. scandens M Hydrangea scandens subsp. liukiuensis 204 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Rokkosan; Kobe. Yamanashi Prefecture: Nishi-yatsushiro. Wakayama Pre- fecture: Nachi. Nara Prefecture: Mt. Omine. RYUKYU ISLANDS. Oki- nawa Island: Kunigawa County. (See Map 6.) Discussion: Hydrangea hirta is the only one of the Asiatic species of Hydrangea which normally does not have sterile flowers. Its leaves have a characteristic shape, which gives them a “nettle-like” appearance. These two characters, lack of sterile flowers and “nettle-like” appearance of the leaves, distinguish this species from H. scandens which it resem- bles most closely in structure of flowers and fruits. Its geographical range is restricted to the island of Honshu in Japan, and Okinawa Island. Hydrangea amagivana Makino (Journal of Japanese Botany 8: 31, 1932) is considered by Makino to represent a possible hybrid between H. hirta and H. scandens subsp. scandens. It is described as having coarsely serrate- dentate leaves, no sterile flowers, and petals cuneate at base. The speci- men which was the basis for Makino’s species came from Izu Peninsula, Honshu Island. A similar plant grows in Nippara, Musashi provinee, ac- cording to Hiyama (see Journal of Japanese Botany 10: 391, 1943). The writer has seen neither of these specimens and therefore cannot evaluate HH. amagiana. 8. Hydrangea scandens (L. f.) Seringe in DeCandolle, Prodromus, 4: 15. 1830. Non Poeppig, 1830, nec Maximowiez, 1867. Shrub or small tree, 1 to 4, or rarely, as much as 6 meters tall; branch- lets and inflorescence pubescent with upwardly appressed hairs, varying in amount from sparse, with hairs 0.1-0.5 mm. long, to more dense with longer hairs, 1-2 mm. long, spreading or appressed, rarely nearly glabrous; leaves ovate, 2-18.5 em. long, 1-8 em. wide, 1.5 to 4 times as long as wide, glabrous to pubescent in vicinity of veins, petiole 1-3 em. long, pubescent as branchlets; inflorescence a branched, compound, flat-topped cluster, 2-18 em. broad, terminal or axillary, involueral bracts none; fertile flowers yellow or blue, few or numerous; hypanthium 0.5-1.5 mm. long; calyx lobes 5, 0.3-3.2 mm. long; petals 5, 1.2-4.5 mm. long, cuneate at base, reflexed, falling with stamens; stamens 10, 1.2-5.5 mm. long, slightly shorter than styles to as much as twice their leneth; ovary half superior; styles usually 3, occasionally 2, 4, or 5, 1.8-3.5 mm. long, united for 0.5-1.5 mm., ty to 2% of their length, free portion 0.5-2 mm. long; capsules 3-8.5 mm. long (including styles), styles 1.5-5.5 mm., united for 0.8-3 mm., less than half their leneth, free portion 0.7—-2.5 mm.; seeds 0.5-0.7 mm. long, 0.3-0.7 mm. wide, oval, blunt at each end; sterile flowers present; sepals (yellow or) white or blue, turning red or yellow in age, 1-3 em. long, nearly as broad as long, margins serrate, apex acute or rounded. (See plate 6, figures 4 and 5.) VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 205 DISTRIBUTION: Western to southwestern China, Formosa, the Philip- pine Islands, Japan, and the Ryukyu Islands. (See Maps 5 and 6.) Kery TO THE SUBSPECIES OF HYDRANGEA SCANDENS 1. Shrubs or small trees up to 4, or even 6 m. tall; inflorescences usually terminal; fertile flowers usually numerous. 2. Stems and leaves pubescent with short upwardly curled hairs or glabrous. c. subsp. chinensis 2. Stems and leaves hirsute pilose. d. subsp. kwangtungensis 1. Small shrubs, usually less than 2 m. tall; inflorescences usually axillary: fertile flowers usually few (10 to 15) in number. 3. Leaves 3.5-9 cm. long, 1-3 cm. wide; bark grayish-brown or tan, dull in ap- pearance. a. subsp. scandens 3. Leaves 2-5 em. long, 1-2 cm. wide, sometimes with white band along midvein; bark reddish-brown, having a somewhat polished appearance. b. subsp. liukiwensis 8a. Hydrangea scandens (lL. f.) Seringe subsp. scandens. Viburnum scandens L. f., Supplementum, 184. 1781. Viburnum virens Thunberg, Flora Japonica, 123. 1784. Hydrangea virens (Thunberg) Siebold, Nova Acta: Academia (caesarea) Leopoldino- Carolina, 14:690. 1829. Hydrangea scandens (lL. f.) Seringe in DeCandolle, Prodromus, 4:15. 1830. Leaves 3.5-9 em. long, 1-3 em. wide, green, upper surface dull, bark grayish brown or tan, dull in appearance. Type LOCALITY: Japan. Type collection: A specimen labeled Viburnum scandens in the J. E. Smith collection of the Linnaean Herbarium in Lon- don is the type of Viburnum scandens L. f. DISTRIBUTION: JAPAN. Honsuu Isuanp: Gifu Prefecture: no locality. Hyogo Prefecture: Nunobiki, Rokkosan. Kanagawa Prefecture: Miyano- shita, Ojigoku, Ohadzu to Hakone, Yokohama, Tonosawa, Hakone. Nara Prefecture: Nara, Mt. Odaigahara, Yoshino. Saitama Prefecture: Mt. Idzugatake, Mt. Takao, Mitsumine-san, Chichibu, Ome. Shizuoka Prefec- ture: Atami, Fujivama, Mt. Amagi on Izu Peninsula, Suruga. Tokyo Prefecture: Nishi-tama-gun, Takao-yama. Wakayama Prefecture: Shingu, Mt. Nachi. Yamanashi Prefecture: Mt. Minobu. Kyusnu Isuanp: Fuku- oka Prefecture: Mt. Kosho. Kagoshima Prefecture: Enoo, Yakushima, Mt. Kirishima, Mt. Takakuma, Kagoshima. Kumamoto Prefecture: Amakusa Island. Miyazaki Prefecture: Sobo-san, Minaminaka-gun. Nagasaki Prefec- ture: Mt. Fugen, Nagasaki. SHikoku Istanp. Tokushima Prefecture: no locality. Kochi Prefecture: Shimonanokama. Formosa. Bankinsing. 8b. Hydrangea scandens (L. f.) Seringe subsp. liukiuensis (Nakai) Me- Clintock, Journal Arnold Arboretum, 37: 375. 1956. 206 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Hydrangea liukiuensis Nakai, Botanical Magazine, Tokyo, 25:63. 1911. Hydrangea luteo-venosa Koidzumi, Botanical Magazine, Tokyo, 39:4. 1925. (No type specimen was designated.) Leaves 2—5 em. long, 1-2 em. wide, variegated with a yellow band along midvein and secondary veins, upper surface lustrous, lower usually glaucous, bark reddish-brown, with more or less polished appearance. (See plate 6, figure 5.) TypE LocaLity of Hydrangea liukiwensis: Ryukyu Islands. Okinawa, Nogodake. Type collection: Tanaka 187 (photo CAS). Specimen not seen by the writer. Distripution: JAPAN. Honsuu Isuanp. Hiroshima Prefecture: Ono- michi. Hyoga Prefecture: Mt. Maya, vicinity of Kobe, Rokkosan. Shizuoka Prefecture: Atami, Amagisan, Shimoda, Shuzenji. Wakayama Prefecture: Mt. Nachi, Koyasan, Yahazuyama. KyusHu IsuANp. Kagoshima Prefec- ture: Makizono. Miyazaki Prefecture: Kitagomura. Nagasaki Prefecture: Nagasaki. Kumamoto Prefecture: Amakusa Island. SHikoKu ISLAND. Mts. of Kochi, Mts. of yo. Kochi Prefecture: Nishinokawa, Yasuda. Ryvu- Kyu ISLANDS. Okinawa Island, Kunigami. Discussion: Hydrangea scandens subsp. liukiuensis is characterized by its small leaves, which often have a yellow band along their midveins, and the reddish-brown stems. It is distinguished from H. scandens subsp. scan- dens, which also occurs in Japan, by these characters. Hydrangea luteo-venosa Koidzumi was described as having small leaves (to + em. long’) which have yellow-white veins and an inflorescence con- sisting of about 20 flowers. The author apparently based it on Asa Gray’s interpretation of the Williams and Morrow specimen collected at Simoda on the Perry Expedition and which Gray determined as H. virens. 8c. Hydrangea scandens (UL. f.) Seringe subsp. chinensis (Maximowicz) McClintock, Journal Arnold Arboretum, 37:3573. 1956. Hydrangea chinensis Maximowicz, Mémoires Académie imperiale des Sciences de St. Pétersbourg, ser. 7, 10(16):7. 1867. Hydrangea Lobbii Maximowicz, 1. c. 15. Type from Philippine Islands, Lobb 446 (isotypes K, E, BM, W, G). Hydrangea Davidii Franchet, Nouvelles Archives Muséum d’histoire naturelle, Paris, ser. 2, 8:227. 1885. Type from Paohing (Moupin), Sikang, David in 1869 (holotype and isotypes P). Hydrangea Pottingeri Prain, Journal Asiatic Society of Bengal, 67:290. 1898. Type from Kachin Hills, Burma, Pottinger, May 4, 1897 (holotype K). Hydrangea Arbostiana Léveillé, Bulletin, Académie internationale de Géographie Botanique, 12:115. 1903. Type from Mont du Collége, Kweichow, Bodinier 1694 (holotype EH, fragment and photo of holotype A, isotypes P, E). Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 207 Hydrangea subintegra Merrill, Philippine Journal of Science, Botany, 3:408. 1908. Type from Batan Island, Philippine Islands, Féniz, June 1907 (isotypes US, K). Hydrangea umbellata Rehder in Plantae Wilsonianae, 1:25. 1911. Type from Kuling, Kiangsi, Wilson 1605 (holotype A, isotypes HE, US). Hydrangea yunnanensis Rehder, 1. c. 37. Two syntypes cited, both from Mengtze, Yunnan, Henry 10236 (A, NY, K), and 10236B (K, EB, S, A, NY). Hydrangea glabrifolia Hayata, Icones Plantarum Formosanarum, 3:106. 1913. Type from Tandaisam, Formosa, Mori, April 1910 (specimen not seen, photo CAS). Hydrangea macrosepala Hayata, I. c. 108. Type from Kokei, Formosa, Nagasawa, April 1903 (specimen not seen, photo CAS). Hydrangea obovatifolia Hayata, 1. c. 109. Type from Terasasan, Formosa, Hayata & Sasaki, July 5, 1912 (specimen not seen, photo US). Hydrangea angustipetala Hayata, Journal College of Science, Imperial University, Tokyo, 30(1):107. 1911. Type from Giran, Formosa, Kawakami & Mori 1385 (specimen not seen, photo CAS). Hydrangea angustifolia Hayata, Icones Plantarum Formosanarum, 1: plate 32. 1911. No description, plate only, which represents the type. Hydrangea angustisepala Hayata, I. c. 2:4. 1912. Type stated to be from Formosa but no specimen was cited. However, the English description is based on the Latin description of H. angustipetala. Hydrangea Davidii Franchet var. Arbostiana Léveillé, Flora Kouy-Tcheou, p. 387. 1914. (Presumably based on H. Arbostiana Léveillé. Publication has not been seen by the author.) Hydrangea pubiramea Merrill, Philippine Journal of Science, Eotany, 12:267. 1917. Type from Tayabas Province, Luzon Island, Philippine Islands, Ramos & Edano herb. no. 26521 (isotypes A, K, US). Hydrangea pubiramea var. parvifolia Merrill, 1. c. 268. Type from Nueva Ecija Province, Luzon Island, Philippine Islands, Ramos & HEdano herb. no. 26511 (isotype US). Hydrangea Kawagoeana Koidzumi, Botanical Magazine, Tokyo, 32:138. 1918. Two syntypes are cited, both from the Ryukyu Islands: Nakanoshima, Kaiwagoe 60, and Takarashima, Ushiwo, May 4, 1917. (Neither specimen seen, photo of Kawagoe 60, CAS.) Hydrangea subferruginea W. W. Smith, Notes from the Royal Botanic Garden, Edinburgh, 11:212. 1919. Type from Htawgaw Valley, Burma, Ward 1542 (holo- type E). Hydrangea chloroleuca Diels, Notizblatt des Botanischen Gartens und Museums, Berlin—Dahlem, 9:1027-1028. 1926. Type from Yungshing Hsien, Kiangsi Prov- ince, China, Hu 799 (isotype A). Hydrangea yayeyamensis Koidzumi, Botanical Magazine, Tokyo, 43:394. 1929. Old- ham 110 (“pro parte’) is the only specimen cited. This specimen, cited previ- ously as a syntype of H. chinensis by Maximowicz, was the basis of H. yayey- amensis, Which was considered by Koidzumi to be a Formosan species. H. chinensis was emended by Koidzumi to include only Chinese plants. Hydrangea formosana Koidzumi, l. c. Five syntypes, all from Formosa, were cited: Oldham 110 “partim” (E, K, P, W, 8S, A, NY), Ford in 1884 (K), and 3 specimens collected by Price (K). 208 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Hydrangea grossiserrata Engler in Engler & Prantl, Nattirlichen Pflanzenfamilien, ed. 2, 18a:207. 1930. Type from Yakushima, Fawrie 3934 (isotype G). Hydrangea stenophylla Merrill & Chun, Sunyatsenia, 1:58. 1930. Type from Lok- chong, Kwangtung, Tso 20617 (NY). Shrubs or small trees, 1 to 4, or rarely, as much as 6 meters tall; leaves 5-18.5 em. long, 1.5-8 em. wide, 1.5 to 4 times as long as wide; inflorescences usually terminal; fertile flowers usually numerous; calyx lobes 0.5-3.2 em. long; capsules 3-8.5 em. long. TYPE LOCALITY: Formosa. Tamsui. Type collection of Hydrangea chi- nensis Maximowiez, lectotype, Oldham 110 (NY), other specimens of this collection, E, K, P, W, S, A. A syntype of H. chinensis Maximowiez is Fortune A42, Chekiang in 1845 (K, G, P, W). DistrIBuTION: Mixed forests and thickets; on Formosa from sea level to 2,500 meters, the Philippine Islands from 700 to 2,500 meters, and from western to southeastern China from 500 to 3,000 meters. (See plate 6, figure 4.) CHINA. ANHwerI: Hwang Shan, Chuhuashan, Tien Chu Shan. CHEKIANG: Mokan Shan, Ping yan, Suian, T’ien-tai Shan, T’ien-mu Shan, Tihtai Shan, Tsinetien, Changhua, Fenghwa, Hangchow, Kewkiang. Fuxiren: Amoy, Ku-dien, Nanping. Hupen: Paohang. Kianest: Tung- huashan, Kuling, Lu Shan, Yung shing hsien. Kwanost: Lingchwan dis- trict, Yao Shan, Ling Wang Shan, Chuen Yuen, Yeo Mar Shan, Hang- on-yuen. KwanatrunG: Lokchong, Tsungfa district, Yao Shan, Yuyuen, Tsing Wan Shan, Kweiyang, Lin fa Shan, Nam Shan, Yam na Shan, Sam Kok Shan, Hong Kong. KwericHow: Kouy-yang, Fan Ching Shan, between Tjiaulou and Baling, near Dayung, Yao ren shan, Pin fa. SIKANG: Kang- tine (Tatsienlu), Hanyuan, Hweili, Paohing (Muping), Tienchuan. SzE- cHUAN: Loshan, He-chang Hsien, Opien, Wa shan, Tahsiangling, Tet- schang in valley Kientschang. YUNNAN: Pingpien, Shunning, Shweli- Salween divide, Tengyueh, Yiliang, headwaters of Shweli River, Salween Valley, Pan pien hsien, Yungshan, Wenshan, Tongtchouan, Shangpa, Kiu- kiang Valley, Mengtze, Ningpo, Tschen Fong chan, Kiao Kia, Gan chouen. BURMA. Kachin Hills, Myitkvina, Kangfang, Kambaiti, Htawgaw, Hpi- maw valley, Maikha-Salween divide. PHILIPPINE ISLANDS. Batan ISLAND: Mt. Ivaya, Santo Domingo de Basco. LrytTE ISLAND: No loeality. Luzon Isuanp: Mountain Province: Benguet Subprovinee, Baguio, Mt. Santo Tomas, Mt. Simacoco, Mt. Pulog, Mt. Pauai, Apayao Subprovinee, Bayag, Bontoe Subprovinee. Tayabas Province: Mt. Dingalan. Nueva Ecija Province: Mt. Umingan. Rizal Province: Balacbac, Montalban, Monte Batay. Bataan Province: Mt. Mariveles. MINDANAO ISLAND. Bukidnon Prov- ince: Mt. Dumaluepihan. Davao Province: Mt. McKinley, Mt. Mayo. Cota- bato Province: Mt. Apo. Misamis Occidental Province: Mt. Malindang. Minporo Isuanp. Mt. Haleon. NeGros ISuAND. Negros Oriental Province: Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 209 Dumaguete, Cuernos de Negros, Canlaon Voleano. Panay Isuanp. Capiz Province: Mt. Madiass, Mt. Macosolon. FORMOSA. Arisan, Bankinsing, Bukiko, Lake Candidius, Chippon-goe in Taitocho, Daksui Valley near Togun, Hassen-zan, Hori, Hokuto, Kapulan, Kelung, Kiirun, Koarong, Siutiam near Banka, Shirisi near Taihoku, South Cape, Sozan, Taipeh, Mt. Chichiseizan, Dakushu, Taihoku, Tomita, Taiheizan, Taitum, Urai, Tamsui. RYUKYU ISLANDS. C.LeEopatra IsuAnp. IrR1omoTE ISLAND, Na- kara River region, Urauchi River, Hidori River. OkINOERABU ISLAND. AKU- SEKI ISLAND. ISHIGAKI ISLAND, between Maezato Yama and peak south of Omato Yama, Omoto-dake. JAPAN. YakusHima ISLAND. Discussion: Hydrangea scandens subsp. chinensis occurs throughout the range of the species with the exception of Japan. It is distinguished from H. scandens subsp. scandens, which occurs in Japan, by its generally larger leaves which appear to be stiff and leathery in comparison with those of subsp. scandens, the somewhat larger size of its individual plants and its usually terminal inflorescences. These differences are usually sufficient to distinguish the two taxa; however, occasional individuals occur in which these differences are not pronounced. Throughout its range Hydrangea scandens subsp. chinensis varies in leaf shape and size. Leaves on the Chinese plants usually are from 5.5—15 em. in length, and 2-6 em. in width, their length being 2 to 3.5 times as ereat as their width, and the pubescence is usually in the vicinity of the midrib and on the petiole. However, there occasionally are found plants with more narrow leaves, ranging from 1.5 to 3.5 em. in width and which are always glabrous. Such a plant is 7'so 20617, which is the type of H. stenophylla Merrill & Chun. In the Philippine plants, the leaves are some- what larger, being up to 18.5 em. long, and 8 em. wide, while in the For- mosan plants the leaves tend to be smaller, being up to 10.5 em. long and 4 em. wide. In addition to the leaf variation in Hydrangea scandens subsp. chinensis, there is some variation in the length of the stamens and their relation to the length of the style, and in the size of the capsules. In the Chinese plants, the stamens are very slightly longer than the styles, to as much as two times as long; the over-all length of the stamens, the greatest for the entire species, is from 1.5 to 5.2 mm., while the length of the styles is not over 3.5 mm. In both the Formosan and the Philippine plants the stamens are equal in length to that of the styles to about one and one-half times their length. In actual measurement the length of the stamens in the Philippine and Formosan plants is about the same, 1.2-4 mm., but in the Philippine plants the styles are somewhat longer (1.8-3.5 mm.) than those of the Formosan (1.5-2.5 mm.). The largest capsules (from 3.5 to 8.5 mm. long) are found in the narrow-leaved Chinese plants, while for 210 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. the Formosan and the broader-leaved Chinese plants the capsules are from 3 to 5.5 mm. long. In the Philippine plants, the capsules are from 5-7.5 mm. long. From the ecollector’s notes occasionally given on the labels there is seen to be a variation in flower color. Both kinds of flowers may be either white (or ? yellow) or blue, and occasionally the sterile flowers are absent. Under Hydrangea chinensis, Maximowiez cited four syntypes, three from China and one from Formosa. The Formosan collection was made by Oldham (No. 110) in 1864 at Tamsui; duplicates of it are available in several herbaria. The specimen in the herbarium of the New York Botan- ical Garden has been selected as the lectotype. The other syntypes cited by Maximowiez are: Fortune A42, collected in 1845 at Amoy, Chekiang Province, and specimens collected by Bunge and Senjawin (both cited by Maximowiez as being in Herb. Fischer of the Botanic Garden in St. Peters- burg), which were not available to the writer. There is some doubt as to whether the Bunge specimen should be identified with Hydrangea scandens subsp. chinensis. Bretschneider (1898) stated that Bunge, who accompa- nied the 11th Ecclesiastical Mission of the Russian Government to Peking, in 1830, made collections in the vicinity of this city. H. scandens subsp. chinensis does not occur in China this far north. The specimen may be H. heteromalla which occurs in the vicinity of Peking. The Seniavin (or Senjawin) specimen is mentioned by Bretschneider (1898, vol. 1, p. 620) as belonging to a collection of Chinese plants said to have been received by Dr. Fiseher from Seniavin ‘who from 1838 to 1851 was Director of the Asiatic Department in the Foreign Office.’ This collection was not made by Seniavin but came to him through the Russian Ecclesiastical Mission at Peking, through whose agency it probably was made in Fukien provinee. Since Fukien is within the range of H. scandens subsp. chinensis the speci- men may belong to this subspecies. Maximowiez based his Hydrangea Lobbw on a eollection by Lobb (No. 446) and said to have been collected in Java. However, according to Merrill (in Philippine Journal of Science, Sect. C, Botany, 10: 185, 1915), “the Kew specimen of this collection is labelled Luzon, and this certainly is correct, as the species is common and widely distributed in the mountains of the northern Philippines.” 8d. Hydrangea scandens (lL. f.) Seringe subsp. kKwangtungensis (Mer- rill) MeClintock, Journal Arnold Arboretum, 37: 373. 1956. Hydrangea kwangtungensis Merrill, Journal Arnold Arborteum, 8:7. 1927. Shrubs 5 to 5 ft. tall; stems, branchlets, and leaves softly villous with hairs 0.5 mm. long; leaves 8-10 em. long, 3-4 em. wide, 2.5 times as long as wide; fertile flowers numerous; calyx lobes 1 mm. long; hypanthium 0.5 mm. long; petals 2.2 mm. long; stamens 1.7-2.2 mm. long; 3 or 4 Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 211 styles, 1.5 mm. long, free portion 1 mm. long; capsules 4.3 mm. long, villous; hypanthium 1.8 mm. long; styles 2.5 mm. long, free portions 1 mm. long, united portion 1.5 mm. long. TyPE Locauity of Hydrangea kwangtungensis: China, Kwangtung Province. Type collection: Peng et al. May 22, 1924, Canton Christian College herb. No. 12017, collected on the National Geographic Society Expedition under the direction of F. W. Wulsin, (holotype US, isotypes mae we. W, BH). DistrisuTion: CHINA. Kwanatune Province: Lung T’au Moun- tain near Ju, Tai Mo Shan, Tapu District. Discussion: Hydrangea scandens subsp. kwangtungensis is distin- euished by the softly villous hairs on its stems and leaves. Its flower and fruit structures are not different from those of the other subspecies of H. scandens. Only two collections of this subspecies were seen, both from Kwangtung Province, indicating that its distribution is doubtless restricted. Subseetion 5. HETEROMALLAE Hydrangea, sect. Euhydrangea, subsect. Heteromallae Rehder, Plantae Wilsonianae, saz, isha. Hydrangea, section 1. Huhydrangea, series 2. Piptopetalae Maximowicz, Mémoires Académie impériale des Sciences de St. Pétersbourg, ser. 7, 10(16):8. 1867. In part. Hydrangea, subgenus Huhydrangea, section Japonico-sinensis, subsection Pipto- petalae (Maximowicz) Schneider, Handbuch der Laubholzkunde, 1:388. 1905. In part. Flowers white; filaments flattened toward their bases; ovary half su- perior; capsule conical at apex; styles shorter than conical disk; seeds caudate. Chinese and Japanese species. 9. Hydrangea paniculata Siebold, Nova Acta: Academia (caesarea) Leo- poldino-Carolina, 14: 691. 1829. Hydrangea Kamienskii Léveillé, Bulietin, Académie internationale de Géographie Botanique, 12:115. 1903. Syntypes from Kweichow: Bodinier 1661 (P, E), and Martin & Bodinier 1661 bis (P). Hydrangea sachalinensis Léveillé, in Fedde’s Repertorium Novarum Specierum, 8:282. 1910. Type from Sakhalin, Faurie 439 (holotype EH, isotypes A, P, BM, W). Hydrangea Schindleri Engler, in Engler & Prantl, Natirlichen Pflanzenfamilien, ed. 2, 18a:203. 1930. In part. Lectotype from Kuling, Kiangsi, Schindler 324 (EH, K, BM, G, W). Much branched shrub or small tree, 0.5-7 m. tall; branchlets and in- florescence pubescent with upwardly appressed hairs 0.3-0.5 mm. long; 212 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. leaves opposite or ternate, ovate, 7-15 em. long, 3-10 em. wide, 1.5 to 2.5 times as long as wide, both surfaces with scattered hairs along veins and veinlets; petiole 1-2.4 em. long; inflorescence a compound, much branched, pyramidal cluster, 7-25 em. long, lowermost branches of inflorescence opening first; bracts numerous, linear to ovate, chartaceous, 1-10 mm. long, 0.2-2.5 mm. wide, subtending main branches of inflorescence, reduced upward, few pili on back and along margins; fertile flowers white, very » numerous; hypanthium 0.8-1 mm. long, glabrous or with few scattered hairs; calyx lobes 5, 0.5-1 mm. long; petals 5, 2-3 mm. long, ovate, re- flexed before falling; stamens 10, 1.7-4.8 mm. long, filaments gradually broadening toward base; ovary half superior; styles 2, 3, or 4, 1.5-2 mm. long, united at base for 0.5-1.5 mm., free for 0.5 to 1.2 mm.; capsule 3.5-5 mm. long, hypanthium 1.5-3 mm. long, styles 2-3.5 mm. long, united for 1.3 to 2 mm., about % to %4 of their length; seeds 1.7-3 mm. long, 0.4-0.5 mm. wide, linear, tapering at each end into minute tail-like ex- tensions; sterile flowers always present, white, 4-lobed, lobes ovate, 1.2-2 em. long, 0.8-1.5 em. wide. (See plate 5, figure 3.) TYPE LOCALITY: Japan. Type collection: the specimens collected in Japan by von Siebold are in the Rijksherbarium in Leiden. Presumably the type of H. paniculata is there, but the writer has not seen it. ILLUSTRATIONS : H. H. Hu & W. Y. Chun. 1933. Icones Plantarum Sinicarum, fase. ITI, pl. 136. Siebold and Zucearini. 1839 or 1840. Flora Japonica, table 61. Gleason, H. A. 1952. Illustrated Flora of the Northeastern United States and Adjacent Canada, 2: 275. DISTRIBUTION: In woods, mixed forests or open places, in southeastern China, from 500 to 1,000 meters, on the islands of Kyushu, Honshu, and Hokkaido in Japan, from sea level to 1,200 meters. (See Map 7.) CHINA. ANHWEI: Chuhuashan, Hwang Shan, Kimen. CHEKIANG: Changshan, Ping Yung, Taishnun, T’ienmu Shan, Tsinetien. FuKIEN: Shaowu, Nanping, Amoy. Hunan: Yun-shan. Kranast: Fencheng, Kiennan, Kuling (in Lu Shan), Lungnan dist., Lu Shan, Ningdu, Wuyuen, Yutu. Kwanasr: Kwei- lin dist., Lingchwan dist., San-Chiang, Tzu-yuan, Yao Shan. KwaNGrTuNe: Linhsien, Lokchong, Meihsien, Yingtak, road to Chang Kiang, Kau Fung, Maan Chi Shan, Yao Shan, Lung T’au near Iu, Fungwan, Yam No Mt., Tsing Leung Shan, Tsing Wan Shan, between Bei Shen and Nan Shung. KweicHow: Kweiting, Kweiyang, Pachai, Tsingchen, Monts de Loutsong- koan, Ganpin, Mantsanping. SrkanG: Paohsing. YUNNAN: Pinfa-Tuchang. JAPAN. Honsuu Isuanp: Nagano Prefecture: Kose, Karuizawa, Shinano. Gumma Prefecture: Yokogawa. Nagano-Gumma_ border, Asama-yama. Aomori Prefecture: Mt. Iwaki, Ikarigaseki, Hakkoda-yama, Tanabu. 213 McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA Vou. XXIX] ‘poulUllojap JOU so.Us0DSeqnd @ ‘9sojyuew0} A ‘jueosoqnd W ‘snoiqe[s Ml: 02,pWouajay nahunuphy -yynjnovund nahuniphyY mm JO VISY U.leysee UT UOTINA! USId *) dVIN BE 214 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Shizuoka Prefecture: Subashiri, Shimoda, Fuji-yama. Nara Prefecture: Odaigahara-san, Omine-sammyaku. Tochigi Prefecture: Mt. Takodate, Lake Chuzenji, Nikko. Kyoto Prefecture: Yase. Gifu Prefecture: No locality. Wakayama Prefecture: Koyasan. Hyoga Prefecture: Ikuno. Kana- gawa Prefecture: Hakone Mts., Yokohama. Yamaguchi Prefecture: Sen- goku. Niigata Prefecture: Mt. Tanigawa, Shimizu. Akita Prefecture: Azuma. Yamagata Prefecture: Yamagata. HokKatpo IstAnp: Oshima Sub- prefecture: Hakodate, Todohokke, Fukayama. Ishikari Subprefecture: Sapporo, Tokobuchi, Otaru. Abashiri Subprefecture: Nokkeushi, Rubeshibe. Nemura Subprefecture: Nishiwada. Kushirokoku Subprefecture: Bokke. Shiribeshi Subprefecture: Otani. KyusHu IsLanp: Kagoshima Prefec- ture: Mt. Kirishima, Kirishima National Park, Sakura-jima, Mt. Onamika, Yakushima. Nagasaki Prefecture: Nagasaki, Unzen. Fukuoda Prefec- ture: Okuma-san. Oita Prefecture: Beppo. SOVIET FEDERATED SO- CIALIST REPUBLIC: Saxwauin: Korsakov, Toronaiba. Discussion: Hydrangea paniculata is distinguished from the other Asiatic species of the section Hydrangea by its pyramidal inflorescence. This type of inflorescence is found elsewhere in H. quercifolia of the east- ern United States. In flower and fruit H. paniculata is most like H. het- eromalla; in fact, in regard to these characters, the two are almost indis- tinguishable, their only difference being the somewhat longer calyx lobes and the somewhat smaller seeds of H. heteromalla. Both have a half su- perior ovary with the styles united for one-half or more of their total length and alternately long and short stamens with filaments broadened toward their bases. H. paniculata is a fairly uniform species, showing no par- ticular variation in any of its parts throughout its geographical range. Hydrangea Schindleri was described by Engler (1930) in his key to the species of Hydrangea. He said of H. Schindleri that it is related to H. chinensis Maximowiez, has long lanceolate leaves constricted in the lower third and narrowly obovate sepals on the sterile flowers, and is found in the Kuling Mountains in Kiangsi, China. According to Rehder (Jour- nal Arnold Arboretum, 14: 202-203, 1933), there were in the Herbarium at Berlin four specimens labeled H. Schindleri, all collected by A. K. Schindler in August and September, 1908, at Lu-Shan in the Kuling Mountains. Duplicates of these specimens were seen by the writer in various European herbaria. Two of the collections, Schindler 325 and 327, are H. scandens subsp. chinensis, and the other two, numbers 322a and 324, are H. paniculata. Engler did not designate a holotype, therefore, Rehder selected number 324 as the lectotype of H. paniculata. 10. Hydrangea heteromalla D. Don, Prodromus Florae Nepalensis, 211. 1825. VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 215 Hydrangea vestita Wallich, Tentamen Florae Napalensis, 2: t. 49. 1826. Type from Nepal, Gosainthan, Wallich Catalogue No. 440/2 (K-W). Hydrangea khasiana Hooker f. & Thomson, Journal Linnaean Society of London, Botany, 2:76. 1858. Type from Assam, Hooker & Thomson August 2, 1850 (holo- type and isotypes K). Hydrangea vestita Wallich var. pubescens Maximowicz, in Mémoires Académie impériale des Sciences de St. Pétersbourg, ser. 7, 10(16): 10. 1867. Two syntypes collected near Pekin by Tatarinow and Kirilow are cited. Not seen by the writer. Hydrangea Bretschneideri Dippel, Handbuch der Laubholzkunde, 3:320. 1893. Named from cultivated material. Hydrangea peckinensis Hort. ex Dippel, 1. c., as synonym of H. Bretschneideri. Hydrangea mandarinorum Diels, in Engler’s Botanische Jahrbiicher fiir Systematik, 29:372. 1900. Type from Nanchuan, Szechuan, Bock & von Rosthorn 1932 (photo and fragment of holotype A). Hydrangea Giraldii Diels, 1. c., p. 373. Five syntypes cited, all collected in Shensi province by Giraldi, Kishan, 1/167, 1171; Tue lian pin, 7/68; Taipai shan, 1169, 1172. Of these the writer has seen only a photo and fragment of No. 1167 (A). Hydrangea xanthoneura Diels, 1. c. Type from Nanchuan, Szechuan. Bock & von Rosthorn 354 (photo and fragment of holotype, isotype W). Hydrangea xanthoneura var. setchuenensis Rehder in Mitteilungen Deutsche Den- drologische Gesellschaft, 21:186. 1912. Type from Wenchwan, Szechuan, Wilson 1323 (holotype A). Hydrangea Bretschneideri Diels var. setchuenensis Rehder, Plantae Wilsonianae, 1:28. 1913. (Based on the preceding. ) Hydrangea Bretschneideri Diels var. lancifolia Rehder, 1. c. Type from Kangting (Tachien-lu), Sikang, Wilson 2412 (holotype A). Hydrangea Bretschneideri Diels var. Giraldii Rehder, I. c., p. 39. (Based on Hydran- gea Giraldii Diels.) Hydrangea xanthoneura Diels var. Wilsonii Rehder, 1. c., p. 27. Type from near Kangting (Tachien-lu), Sikang, Wilson 2407 (holotype A). Hydrangea heteromalla D. Don var. mollis Rehder, 1. c., p. 151. Type from Tsang chan near Tali, Yunnan, Delavay 1148 (holotype P, fragment and photo A). Hydrangea hypoglauca Rehder, 1. c., p. 26. Type from Ichang, Hupeh, Henry 6056 (holotype A, isotypes G, P, E, BM). Hydrangea pubinervis Rehder, 1. c., p. 27. Type from Wa Shan, Szechuan, Wilson 2411 (holotype A, isotypes E, K, BM, W). Hydrangea dumicola W. W. Smith, Notes, Royal Botanic Garden, Edinburgh, 10:39. 1917. Three syntypes are cited from Yunnan: Mingkwong Valley, Forrest 8705 (E, K), 8391 (A, E, K), Tengyueh, Howell 51 (E). Hydrangea macrocarpa Handel-Mazzetti, Anzeiger der mathematisch-naturwissen- schaftlichen Classe: Kaiserliche Akademie der Wissenschaften. Wien, 62:144. 1925. Type from Hweili, Sikang, Handel-Mazzetti 5156 (holotype W, isotypes H, WU). Hydrangea heteromalla D. Don var. parviflora Marquand & Shaw, Journal Linnaean Society of London, Botany, 48:183. 1929. Type from near Gyala, Tibet, Ward 5953 (holotype K). 216 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Hydrangea sungpanensis Handel-Mazzetti, Symbolae Sinicae, 7:444. 1931. Type from Sungpan, Szechuan, Weigold, June-August, 1914 (holotype W). Much branched shrub or small tree, 0.5 to 7 m. tall; branchlets and inflorescence pubescent with upwardly appressed hairs 0.5-1 mm. long; leaves ovate, often broadly so, 8.5-21.5 em. long, 3-14 em. wide, 1.5 to 3 times as long as wide, lower surface pubescent with varying amounts of hairs from sparse to dense enough to cover surface of leaf, or sometimes nearly glabrous with hairs along veins only; upper surface with a few scattered hairs or glabrous; petiole 2.5-4 em. long, pubescent as the stem; inflorescence a rounded, many-branched cluster, 8-80 em. broad; bracts linear to lanceolate, chartaceous, 1-3 em. long, 2-5 mm. wide, few pili on back and along margins; fertile flowers white, numerous; hypanthium 0.7-1.5 mm. long; calyx lobes 5, deltoid, 0.8-3 mm. long; petals 5, 2-3.2 mm. long, 1 mm. wide, ovate, falling separately; stamens 10, rarely 12, 2-6 mm. long, filaments broadened toward base; ovary half superior; styles 3 or 4, 1.5-2 mm. long, united for 0.5-0.8 mm.; capsules 4-5.5 mm. long, hypanthium 1.5-3 mm. long, styles 2-3.2 mm. long, united for 1-2.2 mm., about 14 to °4 of their length; seeds 1.5-2 mm. long, elliptical, tapering at each end into minute tail-like extensions; sterile flowers always present, white, sepals 4, 1-3 em. in diameter. (See plate 5, figure 2.) TYPE LocALIty: Nepal, Gosainthan. Type collection: Wallich in 1821, No. 440/2 of Wallich’s Catalogue. Holotype K-W, a specimen from the Hooker Herbarium at Kew, collected by Wallich in Nepal but without a Catalogue number. Specimens bearing Wallich Catalogue No. 440, from other herbaria (C, P, BM) may be isotypes. ILLUSTRATIONS : H._H. Hu & W. Y. Chun. Icones Plantarum Sinicarum, fase. I, plate 27, 1929, as H. Bretschneideri; and fase. III, plate 137, 1930, as H. xantho- nNeura. N. Wallich. Tentamen Florae Napalensis, t. 49, 1826, as H. vestita. New Flora and Silva, 1: figure 81, 1929, as H. Bretschneideri. DISTRIBUTION: Mixed forests, thickets and open places, west central to northeastern China and the Himalayas. (See Map 7.) CHINA. CHAHaR: Cholu, p.6 CHEKIANG: Tien-mu-shan. HoNaNn: Tsai-Yuan Hsien; Lushih Hsien. HopeH: Peiping, p; Po-hua-shan, p, g; Teche-Ly, Tang Ho, Pei Nin Ting Mt., p; Siao Wu Tai Shan, p; Keang Hsuai Ho, p; Fu-ling-hsien; Tientsin; Ling Shan, p. Hupen: Fanghsien; Hingshan; Ichang, g; Patung; Kienshih, g; Huan Tsao. INNER Moncouis: Wulashan, g. JEHOL: Wel- chang, p; near Hsing Lung Shan, p. Kansu: Choni district (Tao River 6 The small letters, ~, g, t, following the place names refer to the vesture of the lower leaf surface; p—pubescent, t-tomentose, g—glabrous. Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 217 basin), p; Lichen, p;: Old Taochow; Tatung River; Lower Tebbu Country, p, g; Kwan-schan near Lune, g; Sin long chan and Ma _ ho ehan, p. KwanetuNnc: Fuchai. Pineyuan: Tsiyuan. SHansi: Chinyuan, p, g; Kiehsiu, g; Luya Shan, g; Ning-wu, g; Wuchai, p, g; Tanpa district, p; Yaan, p; Hweili, ¢; Yenyuan, ¢. Suryvuan: No locality. SZECHUAN: Chengkow, p; Dzampersheren, p; mountains of Kulu; “Mul Kingdom,” p; Kwanhsien, g; Lifan, p; Maohsien, p; Mt. Mitzuga, p; Nanchwan, p; Mt. Omei, p, g; Opien, g; Mo-tien-ling, g; Tenghsiang-ying, g; Tsien-chuan-fiu, p; Sungpan, p. YUNNAN: Champutong; Chenkang, p; Chungtien, p, t; Mt. Dyinalko, t; Mt. Kaakerpu, t; Lanping, g; Likiang, p, t; Mengtsz, p; Teng-yueh, p; Weisi, p, t; Pe-long-tsin, t; Tong-Tchouan, t; Ma-hong, g; Pe yen tsin, t; Chi-tse-lo, p; Yong-shan, g; Kiukiang Valley, t; Buehhwang Valley; Wenshan; east flank Tali Range, ¢; Tali, t; Tsang chan near Tali, ¢; Minegkwong Valley, p; Mekong-Salween divide, t; Shweli-Salween divide, t, p; Salween River at 27°58’. INDO-CHINA. Tonkin: Chapa. BHUTAN. Singhi Djong, p; Shabjetong, t; Bumthang, p; Trashicho Djong, p; Bela La near Paro, p; Gyetsa to Juto La, t; Paro; Sinchu La, g; Zado ha, g; Fujudin, g; Chapcha, t; Mirchoma; Gyasa Djong, p. BURMA. Hpimaw, p; Adung Valley, p; Muneku Hkyet; Shing Hong Torrent, p; Chawcehi. NEPAL. Mailoong Pati, p; Gosainthan; Jumla, p; Chankheli Lagna, p; Tarakot, Bheri River, t; Sialgarhi, p; Lanetang. TIBET. Cha Gyamda Chu, t; Nayu Chu, p; Kongbo Province, Je Pasum Lake, ¢; Lushu Chu, ¢; Tsangpo Valley, t; gorge of Tsangpo near Gyala, t; Ka Gwr pu Temple, t; Laktang, g; Neawchang Valley, g; above Litang R., p, Yatung, t, Lenda Khala, t; valley of the Seingku, p. INDIA. Assam: Delei Valley, ¢; Khasia Hills, g; Surureem, p; Moflong; Boga Panee, p; Kala Panee. Sikkim: Kaljo- rinie, p; Lachen Valley, p. West BENGAL: Darjeeling, p; Tankra Mt.; Kala- pohri, p; Tumbok. Urrar PrapesH: Kumaun District, Kuthie, p; Gar- hwal, ft. Discussion: Hydrangea heteromalla, unlike the closely related H. pan- iculata, varies throughout its range in leat shape and vesture. The names listed in svnonymy were all based on such variants and none is worthy of recognition. Hydrangea ranthoneura and H. Bretschneideri are prob- ably the best known of these names, but those workers recognizing both of these as constituting separate species had difficulty separating them. Rehder (1924) said of the two that they “are very closely related and both vary with pubescent and glabrescent leaves; the only reliable char- acter to distinguish these two species is apparently the behavior of the bark which in H. Bretschneideri separates in the second year in thin flakes and is of chestnut brown color, without or with few inconspicuous lenticels, while in H. canthoneura the dark adheres to the branch, is conspicuously lenticellate and varies from yellow gray to red brown in color.” However, he separated H. xanthoneura by the glabrous or glabrescent underside of 218 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. the leaves which are more or less cuneate at the base. Of H. Bretschneideri var. lancifolia he said (1911, p. 28), “This variety might possibly be con- sidered a pubescent and narrow leaved form of H. xanthoneura, since it has almost the same kind of bark as var. Wilsonii of that species.” Don based his Hydrangea heteromalla on a specimen collected by Wallich at Gosainthan in Nepal. This Wallich specimen, as in the ease of the one cited for H. anomala Don, is not in the Lambert Herbarium at the British Museum but is in the Wallich Herbarium at Kew. There are three specimens in the Wallich Herbarium written up as follows in Wallich’s Catalogue: “440) Hydrangea vestita Wall. in Herb. 1824. 1. Kumaon, R. B. 2. Napalia 1821 B fimbriata, Napalia, 1821.” Of these three specimens, one of them, 440/2 was collected at Gosain- than by Wallich. This specimen consists of a branch with a full inflores- cence and a small piece of infructescence. The latter does not belong with the inflorescence, but is H. aspera. The flowering branch, with leaves tomentose beneath and more or less rounded, and with entire sterile flowers, agrees with Don’s description of H. heteromalla, which he stated was collected at Gosainthan. Most probably this specimen is the type of H. heteromalla. Of the other two specimens included under No. 440, one, 440/1, was collected in “Kumaon” by Robert Blinkworth. On this sheet are an inflorescence and an infructescence of H. aspera and a small piece of infructescence, which does not belong to the other material, but is H. het- eromalla. The third specimen, 440 8 fimbriata was collected by Wallich in 1821. It is H. aspera. The mixture of these two species in the Wallich material is no doubt in part responsible for the confusion regarding the later mistaken interpretations of H. vestita Wallich, H. heteromalla Don, and H. aspera Don. Hydrangea vestita Wallich was published in the Tentamen on plate 49, without a description. The plate, which is the type for the name, shows a flowering branch (the dominant element of the plate) and a small piece of a fruiting branch. The details of the flower and fruit structure on this plate are not clear, but by comparing this plate with a painting in the East India Company’s collection of paintings in Herb. Kew, it be- comes evident that two elements are represented in the original painting, and that the plate was adapted from it. The dominant element, both in the painting and the plate in the Tentamen, a branch with a full inflor- escence, is surely what Wallich intended his H. vestita to be. The lesser element, the piece of infructescence, is H. aspera. Comparison of this plate with the Wallich specimen No. 440/2, described above as consisting of a full inflorescence of H. heteromalla and a small intructescence of H. aspera, VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 219 shows that the plate with its mixture of the same two species was made from the specimen. It may be concluded that this specimen served as the basis of H. heteromalla Don and later for the plate of H. vestita Wallich. The material on which Hydrangea khasiana is based was collected by Hooker and Thomson in the Khasia Mountains at Surureem in the valley of the Kala-panee River. Hooker and Thomson said of it that it differs from “H. vestita,” that is H. heteromalla, by its more robust habit and its broader leaves, which are cordate at the base. H. heteromalla varies in habit from a shrub to a small tree. Its leaves are also variable. The variation shown by H. khasiana is not significant. Maximowiez (1867) based Hydrangea vestita var. pubescens on two specimens collected by Tatarinow and Kirilow in the Peking area of north- ern China. The collections of both of these men were eventually deposited in the herbarium of the Botanic Garden in St. Petersburg, where they were available to Maximowiez for study. Maximowiez stated that his proposed variety was conspecific with the Himalayan H. vestita (referred in this paper to H. heteromalla) and he followed Hooker and Thomson in using the name H. vestita instead of H. heteromallaa for the Himalayan plant. Since he stated’ that he saw the Wallich specimen from Gosainthan on which H. vestita Wallich was based, there is no doubt that he used the name H. vestita var. pubescens for the species here referred to H. heteromalla. However, he took the varietal epithet from H. pubescens Decaisne (Flore des Serres, 4: t. 378, 1848). He stated that this name, known to him only from the illustration, was not different from his proposed variety. Decaisne’s H. pubescens, however, was described from a cultivated plant grown from seed which was said to have come from Japan. The illustration represent- ing this plant does not show any detail of flower structure and no speci- men is cited in the description. Therefore, it is not possible to identify the plate, other than to say that if the plant shown were grown from seed of Japanese origin it would not be the same as Maximowicz’s H. vestita var. pubescens, which is based on the two collections from Peking. It seems best, therefore, to regard the name H. pubescens Decaisne as a nomen dubwum. Hydrangea Bretschneideri was named by Dr. Leopold Dippel, who was director of the botanical garden in Darmstadt, Germany. Dippel stated, in his description of this plant, that it was grown from seed sent from Peking by Dr. Bretschneider, that it had been referred to by the names H. vestita and H. peckinensis, and that it was sometimes confused with H. pubescens Decaisne. Dippel did not cite a specimen in his de- scription of H. Bretschneideri; hence no type specimen was designated. 7 See Maximowicz (1867, p. 10): ‘‘Ipsi vidi e Kumaon, Gossainsthan (ex hb. Lambert) et spec. Wallich No. 440, flor. et frf.” 220 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. However, cultivated specimens grown in the garden at Darmstadt have been distributed to various herbaria (Koehne 470 and Purpus June 15, 1924 are both in the herbarium of the Arnold Arboretum.) These specimens labeled “Hydrangea bretschneideri Dippel” doubtless represent the species deseribed by Dippel. Subsection 6. MAcROPHYLLAE Hydrangea, sect. Hydrangea, subsect. Macrophyllae McClintock, Journal Arnold Arboretum, 37:374. 1956. Hydrangea, sect. Huhydrangea, series 2. Piptopetalae Maximowicz, Mémoires Acad- émie impériale des Sciences de St. Pétersbourg, ser. 7, 10(16): 8 (Revisio Hydrangearum Asiae Orientalis). 1867. In part. Map 8. Distribution in Japan of @ Hydrangea macrophylla subsp. serrata A Hydrangea macrophylla subsp. macrophylla VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 221 Hydrangea, subgenus Euhydrangea, sect. Japonico-Sinensis, subsect. Piptopetalae (Maximowicz) Schneider, Handbuch der Laubholzkunde, 388. 1905. In part. Flowers white, blue, or pink; ovary inferior in anthesis, half superior in fruit; capsule conical at apex; styles about same length as conical disk; seeds caudate. A single species, Hydrangea macrophylla (Thunberg) Seringe, widespread in eastern Asia. 11. Hydrangea macrophylla (Thunberg) Seringe in DeCandolle, Pro- dromus, 4:15. 1830. Viburnum macrophyllum Thunberg, Flora Japonica, 125. 1784. Shrubs 1-3 m. tall; branchlets and inflorescence glabrous or pubescent with upwardly curled hairs; leaves 5.5-19 em. long, 2.5-13 em. wide, their leneth about 1.5 to 4 times their width, ovate or broadly ovate to more or less obovate, texture varying from thin to thick; petioles 1-4 em. long; bracts none; inflorescence a flat-topped, compound, several-branched clus- ter; fertile flowers white; hypanthium 1.2-2 mm. long; ealyx lobes 5, deltoid, 0.5-1 mm. long; petals 5, 2-3 mm. long, reflexed before falling; stamens 10, 2-6 mm. long; styles usually 3, occasionally 2, 4 or 5, 1-2 mm. long, in early anthesis free to base, ovary then inferior; styles in late an- thesis united for 0.3-1 mm. from base, ovary then half superior; base of ovary and capsule, sometimes tapered toward pedicel; capsule (ineclud- ing styles) 2.5-8 mm. long, hypanthium 1.5-5 mm. long, styles 1.2-3.5 mm. long, free portion of styles 0.7-1.5 mm., united portion of styles 0.5—2 mm. long; seeds 0.5-0.7 mm. long; sterile flowers white or blue; petaloid sepals 1-3 em. across, entire or serrate. (See plate 6, figures 2 and 3.) DISTRIBUTION: The eastern Himalayas, southern China, and Japan. (Map 8 shows distribution in Japan.) Key TO THE SUBSPECIES OF HyDRANGEA MACROPHYLLA 1. Stems and leaves hispid with spreading hairs 1-3 mm. long. d. subsp. Chungii 1. Stems and leaves glabrous or pubescent with hairs less than 1 mm. long. 2. Petals 3-4 mm. long; styles in anthesis 2 mm. long; capsules 6-8 mm. long; leaf length about one and one-half times greater than width. a. subsp. macrophylla 2. Petals 2-3 mm. long; styles in anthesis 1-2 mm. long; capsules 2.5-6 mm. long; leaf length one and one-half to four times greater than width. 3. Capsules 2.5-5 mm. long, with styles 1.2-2.5 mm. long, free portions 0.7-1.5 mm. long. (Planis of Japan.) b. subsp. serrata 3. Capsules 4-6 mm. long, with styles 2-4 mm. long, free portions 1.5—2 mm. long. (Plants of the eastern Himalayas and southern China.) c. subsp. stylosa lla. Hydrangea macrophylla subsp. macrophylla. Hydrangea macrophylla (Thunberg) Seringe in DeCandolle, Prodromus, 4:15. 1830. 222 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Viburnum macrophyllum Thunberg, Flora Japonica, 125. 1784. Hydrangea macrophylla var. normalis Wilson, Journal Arnold Arboretum, 4:234. 1923. Type from Japan, Suzuki July 1917 (holotype A). Hydrangea maritima Haworth-Booth in The Hydrangeas, p. 38. 1950. (Based on Hydrangea macrophylla var. normalis.) Branchlets, inflorescence, and leaves glabrous; leaves 7.5-19 em. long, 4-13 em. wide, their length about 1.5 times their width, broadly ovate to more or less obovate, margins coarsely serrate, texture “thick”; petioles 4 em. long; petals 3.5-4 mm. long; stamens 3.5-4.5 mm. long; styles 3, 2 mm. long in early anthesis; capsules (including styles) 6.5-8 mm. long, hypanthium 3.5-5 mm. long, styles 2.5-3.5 mm. long, united portion of styles 1.5-2 mm. long, free portion of styles 1-1.5 mm. long. DISTRIBUTION: Japan. Chiba (Boso or Boshu) Peninsula on the coast of east central Honshu and the islands of Oshima (De Vries) and Hachijo, from sea level to 150 meters. TYPE LocaLity of Viburnum macrophyllum: Japan. Type collection: four syntypes collected by Thunberg. According to Hara (1955) two were collected in Japan and two were collected in Thunbere’s garden in Uppsala. All specimens belong to a garden form having nearly all sterile flowers in its inflorescences. SPECIMENS EXAMINED: JAPAN. Honsuu Isuanp. Tokyo Prefecture: Hachijo Island, Oshima, Chiba (Boso or Boshu) Peninsula. lib. Hydrangea macrophylla (Thunberg) Seringe subsp. serrata (Thun- berg) Makino in Journal of Japanese Botany, 6(7) :11. 1929. Viburnum serratum Thunberg, Flora Japonica, 124. 1784. Hydrangea serrata (Thunberg) Seringe in DeCandolle, Prodromus, 4:15. 1830. Viburnum cuspidatum Thunberg, I. c. 125. Type from Japan, collected by Thunberg (UPS). Hydrangea cuspidata (Thunberg) Miquel, Annales Museum Botanicum Lugduno- Batavum, 3:98. 1867. Hydrangea yesoensis Koidzumi, Botanical Magazine, Tokyo, 40:347. 1926. Type from Hakodate, Hokkaido, Maximowicz in 1861 (isotypes US, K, S, W, NY). Branchlets, leaves, and inflorescence glabrous or pubescent with up- wardly appressed hairs; leaves 5.5-15 em. long, 3-6 em. wide, their length 1.5 to 3 times their width, ovate, acute or acuminate at apex; petioles 1—3 em. long; hypanthium 1.2-1.3 mm. long; calyx lobes 5, 0.5—0.6 mm. long; petals 2-3 mm. long; stamens 2-5 mm. long; styles usually 3, occasionally 2 or 4, 1-1.7 mm. long in early anthesis; styles united for 0.3-1 mm. from base in late anthesis; capsule (including styles) 2.5-5 (rarely 6) mm. long, united for 0.5-1.5 mm., free portion of styles 0.7-1.5 mm. long. Vou. XXIX|] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 223 Type LOCALITY of Viburnum serratum: Japan. Type collection: five syntypes in the Thunberg Herbarium, specimen 8 selected as lectotype by Wana (41955). DISTRIBUTION: Japan and southern Korea, in woods and forests usually in mountainous areas, at elevations of 70 to 1,500 meters. JAPAN. Hon- sHu Isuanp. Hyogo Prefecture: Arima. Aichi Prefecture: Shinano. Gifu Prefecture: no locality. Gumma Prefecture: Ikaho, Mt. Haruna. Tochigi Prefecture: Nikko to Lake Chuzenji, Lake Chuzenji, Nikko. Ibaraki Pre- fecture: Mt. Kaba. Kanagawa Prefecture: Hakone Mts., Manazuru, Nagoa, Miyanoshita, Yokohama, Sagami Bay. Kyoto Prefecture: Ohara, Kurama- yama. Wakayama Prefecture: Koya. Yamagata Prefecture: Ubayu. Ya- manashi Prefecture: Lake Shoji, Jizoga-take, Nambu. Nara Prefecture: Mt. Omine. Nagano Prefecture: Karuizawa, Kisokoma-ga-take, Norikura, Agawa, Kawagishi, Otake River, Asama-yama. Shizuoka Prefecture : Shojiko, Shimoda. Saitama Prefecture: Mt. Buko, Mitake, izuga-take, Hikawa, Shi- mura, Mt. Mitsumine. Niigata Prefecture: Mt. Tanigawa, Sado Island. Osaka Prefecture: Mt. Kongosan. Akita Prefecture: no locality. Tokyo Prefecture: Tokyo. Aomori Prefecture: Hokkadayama, Lake Towada, Hirosaki. KyusHu Isuanp: Nagasaki Prefecture: Tsu-shima. Kagoshima Prefecture: Mt. Kirishima. Oita Prefecture: Kujisan. Fukuoka Prefec- ture: Mt. Ukidake. SHikoku ISLAND: Kochi Prefecture: Shimonanokawa, Takaoka, Shiraga-yama, Engyoji. Hoxkamo IsuANp: Oshima Subprefec- ture: Hakodate, Ichinowatari. Ishikari Subprefecture: Garugawa, Moiwa- dake, Teine-yama. Shiribeshi Subprefecture: Shiribeshi san. KOREA. QUELPAERT ISLAND. lle. Hydrangea macrophylla (Thunberg) Seringe subsp. stylosa (Hook- er f. & Thomson) McClintock, Journal Arnold Arboretum, 37: 373. 1956. Hydrangea stylosa Hooker f. & Thomson, Journal Linnaean Society of London, Botany, 2:75. November 1857 (holotype K). Hydrangea cyanema Nuttall in Curtis’s Botanical Magazine, t. 5038. March 1858. Described from cultivated material. Hydrangea taroensis Handel-Mazzetti, Anzeiger der mathematisch-naturwissen- schaftlichen Classe: Kaiserliche Akademie der Wissenschaften. Wien, 62:144. 1925. Type from near Tibet-Burma border, Irrawadi River, Handel-Mazzetti 9455 (holotype W, isotypes A, E). Hydrangea kwangsiensis Hu, Journal Arnold Arboretum, 12:152. 1931. Type from Kwangsi, Ching 5836 (photos and fragments of holotype A, NY, isotype W). Hydrangea indochinensis Merrill, Journal Arnold Arboretum, 23:167. 1942. Type from Chapa, Indochina, Petelot 2687 (holotype A). Branchlets, leaves and inflorescences glabrous or pubescent with up- wardly curled hairs; leaves ovate, 7.5-14.5 em. long, 2.5-5.5 em. wide, length 224 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. two to four times as great as width; leaf margins finely to coarsely ser- rate or entire; petioles 0.7-3 em. long; fertile flowers lavender or blue; petals 2 mm. long; stamens 4-6 mm. long; styles 3, 4, or 5, 1.5 mm. long; capsules (ineluding styles) 5-6 mm. long, hypanthium 2-3 mm. long, styles 2-3 mm. long, united for 0.5-1.5 mm., free portion of styles 1.5—-2 mm. Type Locauity of Hydrangea stylosa Hooker f. & Thomson: India, Sikkim, Lachen (8,000 ft.). Type collection: Hooker, August 4, 1849 (holotype K). ILLUSTRATIONS : H. H. Hu and W. Y. Chun. 1933. Icones Plantarum Sinicarum, fase. III, pl. 135, as H. kwangsiensis Hu. Curtis’s Botanical Magazine, 1858. Table 5038, as H. cyanema Nuttall. DISTRIBUTION: Forests and wooded areas in the eastern Himalayas of Bhutan and Sikkim and mountains of Yunnan, Kwangsi, Kweichow, Kwanetung and Indochina, at elevations of 500 to 3,300 meters. CHINA. Kwanasi: Chu-feng Shan, north of Luchen Hsien on border of Kweichow, Ling Wang Shan. Kwanetune: Yao Shan. KweicHow: Yao-ren-shan, San hoa. YUNNAN: Eastern flank of Tali Range; Yangpi Valley; western flank of Tali Range, Shweli-Salween divide; Tali Range at middle part of Chung Ho Mt.; Shweli River drainage basin to summit of Shweli-Salween water- shed east of Tengueh; Shunning Hsien; Tali Hsien. BURMA. Irrawadi River, near Tibet-Burma border. BHUTAN. Tsanka to Chendehi; Trashi- yangsi Chu; near Tongsa; Ridang near Tongsa; Rudong La; Kancham. INDIA. Sikkim: Talung, Lachen. INDOCHINA. Vietnam: Chapa. Discussion: Hydrangea cyanema Nuttall ex Hooker was described and illustrated in Curtis’s Botanical Magazine, plate 5038, in 1858. Hooker stated that this plant was “introduced by Mr. Nuttall from Bhotan where it was discovered by Mr. Booth.” Thomas J. Booth collected rhododen- drons and other plants in the eastern Himalayas for his unele, Thomas Nuttall of Nutgrove, Rainhill, Prescott. From Booth’s seeds, a number of rhododendrons were grown and introduced into cultivation for the first time. The area of ‘‘Bhotan” where Booth collected is to the east of the present Bhutan, and appears to be in the Daphla (Dafla) and Aka Hills (see Stapf in Curtis’s Botanical Magazine, t. 8078). In Herb. Kew is a fruiting specimen collected by Booth in “Bootan” and which, accord- ing to the label, was from “Herb. Nuttall.” It probably was this collection that furnished the seed from which was grown the plant used for the illustration of Hydrangea cyanema. In the description which accompanies. the illustration, no specimen is cited; therefore, the illustration may be considered to represent the type. The specimen on which Hydrangea kwangsiensis was based, Ching 5836 Vou. XXIX| McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 225 from Kiangsi, and a second specimen, T'siang 6285 from Kweichow, differ slightly in leaf margin and leaf texture from the other material included in subsp. stylosa. In most of the material of this subspecies the leaves are serrate and “thin” in texture, but the two specimens from Kiangsi and Kweichow have leaves which have entire margins and are somewhat “thick” in texture. These leaf differences are of a similar nature, but not as pro- nounced, as the leaf differences between the Japanese subsp. serrata and subsp. macrophylla. Probably because of the small size of the fertile flowers, Hydrangea macrophylla subsp. stylosa has been confused with other species in the Himalayan region and in western China. For the same reason, perhaps, its relationship to H. macrophylla has been overlooked. It is of interest that this relationship extends the geographical range of H. macrophylla from the Himalayas through China and to Japan. lid. Hydrangea macrophylla (Thunberg) Seringe subsp. Chungii (Reh- der) McClintock, Journal Arnold Arboretum, 37: 374. 1956. Hydrangea Chungii Rehder, Journal Arnold Arboretum, 12:69. 1931. Shrubs 1 to 2 m. tall; stems, branchlets and leaves hirsute with spread- ing hairs, 1.5-3 mm. long, in addition shorter, upwardly curled hairs along stem, 0.5 mm. long; leaves 10-19 em. long, 5-8 em. wide, 2 to 2.5 times as long as wide, petioles 1-4 em. long; mature fertile flowers not seen; styles (in buds) 3 or 4, free to their bases; capsules 3 mm. long; hypan- thium 1.8 mm. long; styles 1.2 mm. long, free portions 0.8 mm. long, united portions 0.4 mm. long. TYPE LOCALITY: Yenping, Fukien. Type collection: Chung 3043 (holo- type A, isotypes A, UC, K, W). ILLUSTRATIONS: H. H. Hu and W. Y. Chun. 1933. Icones Plantarum Sinicarum, fase. II, plate 184, as Hydrangea Chungu. DISTRIBUTION: Central Fukien, China. CHINA. Fukien: Yenping. Discussion: Hydrangea macrophylla subsp. Chung is characterized by its hirsute stems and leaves. Of the two collections from Yenping one is in fruit and the other in flower. The flowers of this latter specimen are immature but because the styles are free to the base the plant is consid- ered to belong to H. macrophylla. Known only from the three collections cited above the subspecies is doubtless restricted in its range. 226 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Discussion oF Hydrangea macrophylla AND THE ORIGIN oF Its GARDEN ForMsS Hydrangea macrophylla is probably the best known species of the genus because to it belong the many cultivated forms of the garden hydrangea. These forms with their sterile-flowered inflorescences represent an unusual condition of subsp. macrophylla found on the east central coast of the island of Honshu. Hydrangea macrophylla and H. serrata, long consid- ered to be closely related Japanese species, were combined by Makino (1929) into a single species. Hara (1955) reviewed the nomenclatural problems involved in this combination. The two species from the Asiatic mainland, H. stylosa and H. Chungu, are combined here for the first time as subspecies of H. macrophylla. The relationship between subsp. stylosa and subsp. serrata is close and it is difficult to distinguish between the two. Subsp. macrophylla may be separated from the others by its larger flower parts and thick-textured leaves, and subsp. Chungii is separable by its hispid spreading hairs. Within recent years the question of the origin of the garden forms of Hydrangea macrophylla has been discussed from two different points of view. Wilson (1923) suggested that the commonly cultivated form with the globose inflorescences consisting of all sterile flowers “is simply an anomalous condition of a littoral plant very common on the shores of Boshu peninsula in central Hondo, om Oshima or De Vries’ Island, on Hachijo and others of the seven isles of Idzu, described in detail . . . under the name of Hydrangea macrophylla var. normalis.” Wilson goes on to say that the ‘replacing of the ordinary flowers by sterile flowers having petaloid sepals occurs in several species of Hydrangea .. . it is probable that this character can and does appear as a sport in many species of Hy- drangea.”’ In 1917, Wilson found on Oshima a plant which he considered to be a wild prototype of the garden plant and which was no different from the plant which he knew so well, both at home and in China and Japan. Al- though the plant which Wilson found on the coastal area of central Honshu was growing not far from Yokohama and Tokyo, its relation to the cultivated hydrangeas had not been suggested until Wilson made his observations. Wilson recognized his coastal plant as being different from Hydrangea serrata, the Japanese species most closely related to H. macrophylla, which he knew as a woodland plant growing in the mountains throughout Japan. Differences in the leaves between these two species, not readily seen in herbarium material, were pointed out by Wilson. He distinguished between the two in the field on the basis of their leaf textures and the greater vigor of H. macrophylla. The taxonomy of H. macrophylla and H. serrata has been complicated by the similarity between the two and by the cultivation of numerous garden forms of both. This similarity is so Vou. XXIX| McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 227 strong that Japanese botanists, particularly Makino and Hara, consider them to be coastal and inland forms of the same species. Haworth-Booth (1950) discussed the origin of the garden hydrangeas from the point of view of their being hybrids. He starts with the premise that there are three species from the woodland areas of Japan for which he used the following names: H. acuminata, H. japonica, and H. Thun- bergu. He characterized these taxa chiefly on the basis of their habit and cultural responses, but also mentioned the shapes of the capsules and the color and margins of the enlarged sepals of the sterile flowers. Ac- cording to him H. acuminata has capsules tapered to their bases and entire sepals which are “pink in neutral or limy soil and blue in acid soil”; H. japonica has capsules almost round and serrated sepals which “open white and then, absolutely irrespective of whether the plant is growing in neutral to limy or acid soil, turn crimson owing to the effect of sunlight”; H. Thunbergu, which is “earlier flowering and much more dwarf in growth with exceptionally numerous, tiny, orbicular ray-flowers of unusually vivid pink and reddish shoots (Haworth-Booth, 1950, p. 20). It is Haworth-Booth’s contention that these three so-called species which share ‘“‘certain characters such as woodland habit, the quality of flowering from all of the side shoots, and leaves that are thinner” in texture than the maritime form of H. macrophylla called var. normalis by Wilson, have hybridized to produce the plant which Thunberg described as Viburnum serratum and is here called H. macrophylla subsp. serrata. He further contended that these three woodland species, in addition to producing hybrids among themselves, have produced hybrids with the maritime plant described by Wilson under the name H. macrophylla var. normalis, and that from these hybrids have originated the plant which Thunberg described as Viburnum macrophyllum and which is here called H. macrophylla subsp. macrophylla. Haworth-Booth gave the name H. maritima to Wilson’s H. macrophylla var. normalis because he considered Thunbere’s name (Viburnum macrophyllum) to have been applied to a “hybrid race,” and therefore, this epithet “cannot be used for a wild species which is only one of the ancestors of that hybrid race” (Haworth- Booth, p. 38). From an examination of the available herbarium material and observa- tions on living cultivated plants, it appears that the garden forms of Hydrangea macrophylla subsp. macrophylla are what Wilson said they were: an anomalous condition of the plant which he found in the coast of central Honshu. There is no evidence, morphological or otherwise, to substantiate Haworth-Booth’s suggestion of there being three inland mon- tane species of Hydrangea in Japan. He based his three “species” on plants named by Siebold and Zuecarini. The characters which Haworth- Booth ascribed to them were for the most part ecological and cultural 228 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. responses, such as tolerance to shade, flower color, time of flowering, habit of flowering, and texture of leaves. It is apparent that these are not morphological characters upon which species can be based. Doubtless, through mutation a plant or plants having all sterile flowers in their inflorescences appeared in the wild populations of both the montane and the maritime taxa, subsp. serrata and subsp. macrophylla respectively, such plants were selected for cultivation and from them have come the various garden forms. The following names have been given to specimens of Hydrangea pre- sumably of garden origin. From the descriptions and/or illustrations the names apply for the most part to H. macrophylla subsp. macrophylla. Hortensia opuloides Lamarck, Eneyeclopédie Méthodique, Botanique, 3: 136, 1789. Based on a plant cultivated on Mauritius (Ile de France) and probably sent to the Jardin des Plantes in France by Commerson. Primula mutabilis Loureiro, Flora Cochinchinensis 104, 1790. Mer- rill in his “Commentary on Loureiro’s Flora Cochinchinensis” (Transactions American Philosophical Society, 24(2):177, 1935) says “the description apples unmistakably to the commonly cultivated garden Hydrangea.” Hortensia japonica Gmelin in Linnaeus, Systema Natura, ed. 13, I, pt. 1, 722, 1791. Name only. Generic description based on that of Jussieu. Hydrangea hortensis Smith, Iecones Pictae Plantarum Rariorum, p. 12, t. 12, 1792. Through the ageney of Sir Joseph Banks a living hydrangea was brought to Kew in 1789 from China. This plant was named Hydran- gea hortensis by Sir James Smith who recognized it as being related to H. arborescens Linnaeus. Hortensia mutabilis Schneevoogt & Sehwegman, Icones Plantarum Rariorum, t. 36, 1793. The illustration shows an inflorescence consisting entirely of sterile flowers which are violet or light purple in color. The dissections are not quite accurate but the plate surely represents a garden form of Hydrangea macrophylla. Hydrangea Hortensia Siebold, Nova Acta: Academia (caesarea) Leo- poldino-Carolina, 14(2) : 688, 1829. Under this name Siebold cited as syno- nyms Hydrangea hortensis Smith, Hortensia opuloides Lamarck, and Hor- tensia speciosa Persoon. The plant, he stated, was cultivated in Japan, but he did not illustrate it in the Flora Japonica, perhaps because it was so well known. Hydrangea japonica Siebold, 1. c. 689. Siebold & Zueearini, Flora Japonica 1: 106, t. 53, 1839 or 1840. Plant illustrated has pink fertile and sterile flowers. Siebold said of Hydrangea japonica that it was a garden plant cultivated for centuries in Japan but found in the wild state in the mountains of that country. He described it as having pink flowers, and Vou. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 229 as being robust and large in all parts. The illustration shows it as being glabrous, this together with its robust habit suggests H. macrophylla subsp. macrophylla rather than H. macrophylla subsp. serrata, despite the fact that Siebold said that the plant grows wild in the mountains of Japan. Wilson cited this name as a synonym of his H. macrophylla var. normalis f. rosea, which he said “has pink, more or less toothed ray-flowers, other- wise it is very similar to the wild form.” Hydrangea japonica, Botanical Register, 30: t. 61, 1844. Lindley ered- ited this name to Siebold and cited t. 53 of Siebold and Zueearini’s Flora Japonca, The Botanical Register plate shows a plant having an inflores- cence in which the fertile flowers are rose colored and are surrounded by white sterile flowers. In flower color, therefore, it differs from Siebold and Zucearini’s plate in which the flowers are pink. The two probably repre- sent color forms of Hydrangea macrophylla subsp. macrophylla. Hydrangea japonica var. caerulea, Curtis’s Botanical Magazine, t. 4253, 1846. This plate shows a plant having an inflorescence in which the fertile flowers are blue and the sterile flowers white. W. J. Hooker, who gave this blue-flowered plant the varietal name listed here, said that it is more handsome than the pink and rose flowered plants pictured in Flora Japonica and Botanical Register. It is undoubtedly a eolor form of Hydrangea macrophylla subsp. macrophylla. Hydrangea Sitsitan Siebold, Nova Acta: Academia (caesarea) Leo- poldino-Carolina, 692, 1829. Based on garden material, Siebold considered Hydrangea Sitsitan to be a dubious species, perhaps a form of his A. hortensis. Hydrangea Thunbergu Siebold, |. c., 690. Siebold & Zucearini, Flora Japonica, 1: 111, t. 58, 1839 or 1840. Siebold cited Viburnum serratum Thunberg as a synonym of his Hydrangea Thunbergu but said that be- cause Thunbere’s description was incomplete and imperfect he was not certain whether his plant was the same. He said further that the plant was known to him only in cultivation although it was said to grow in the provinees of Awa and Sanuki on Shikoku and in the mountains of Honshu. From the illustration it appears that this plant may be H. macrophylla subsp. serrata. Hydrangea acuminata Siebold & Zucearini, Flora Japoniea, 1: 110, t. 56, 1839 or 1840. The illustration shows a plant with blue flowers. Like /. Thunbergu, it is difficult to tell from the illustration whether the plant represents H. macrophylla subsp. serrata. However, since it was said to grow in the mountains of Honshu and Kyushu it may represent H. macro- phylla subsp. serrata rather than H. macrophylla subsp. macrophylla. Table 57 of the Flora Japonica, said also to represent Hydrangea acuminata, 230 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. shows a plant with capsules. The hypanthium of the capsule is somewhat abruptly constricted at the pedicel suggesting that the illustration should be referred to H. scandens subsp. scandens rather than H. macrophylla subsp. serrata. Hydrangea Azisai Siebold & Zucearini, Flora Japonica, 1: 104, t. 51, 1839 or 1840. The plant illustrated has blue fertile flowers and several large pale blue sterile flowers, with some pubescence on the peduncles and pedicels. This name is based on a Japanese garden plant which Siebold and Zucearini said was much sought after as an ornament in gardens and eroves around temples. They said that it came originally from China and is so similar to the “Hortensia” that one can regard it as having had the same origin. Hydrangea Belzonii Siebold & Zucecarini, |. c., 109, t. 55. Siebold and Zucearini described Hydrangea Belzoni as frequently cultivated in Japanese gardens, and having two forms, one with all sterile flowers and the other with both fertile and sterile flowers. The flowers in both are blue. They distinguished it from other garden forms of Hydrangea macro- phylla subsp. macrophylla by its verticillate as well as opposite leaves. Hydrangea Burgeri Siebold & Zucearini, 1. c., 111, t. 57. This name was used by Siebold and Zucearini for a Japanese garden plant with both sterile and fertile flowers. It appears from the illustration to be H. macro- phylla subsp. serrata. Hydrangea Otaksa Siebold & Zucearini, |. c., 105, t. 52. The illustra- tion of Hydrangea Otaksa shows a plant with all sterile flowers and with leaves which appear to be thick in texture, as in H. macrophylla subsp. macrophylla. Siebold and Zucearini considered this plant to strongly resem- ble Siebold’s H. hortensia but preferred to keep it separate because of its obovate leaves and large clusters of blue flowers. It was, they said, rare in Japan and probably just recently imported from China. They also said that it was cultivated under the name of “Otaksa” in the botanical garden on the Isle of Deshima. It was on this small island in Nagasaki Bay that Thunberg and later Siebold stayed when they were in Japan and from whence came many of the plants which they both described. Hydrangea stellata Siebold & Zucearini, |. c., 112, t. 59. H. stellata is described and illustrated as having mostly sterile flowers which are double and blue but not numerous. There is no clue as to which form this represents of Hydrangea macrophylla. Hydrangea japonica rosalba Van Houtte, Flore des Serres, 16:75, tt. 1649, 1650, 1865-1867. Of the two illustrations representing Hydrangea japonica rosalba one has an inflorescence having white sterile flowers and the other an inflorescence with rose-colored sterile flowers. Both have numerous VoL. XXIX] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 231 white fertile flowers surrounded by the sterile flowers and thick appearing leaves which suggest H. macrophylla subsp. macrophylla. Hydrangea Lindleyana Hort. ex Lavallée, Arboretum Segrezianum, 124, 1877. This horticultural name was doubtless applied to one of the garden forms of H. macrophylla. Wavallée lists it as a synonym of H. Otaksa Siebold & Zucearini in his list of plants growing in the garden at Segrez in France. Hydrangea sinensis Hort. ex Lavallée, 1. c., 128. Lavallée listed this horticultural name as a synonym under H. japonica Siebold & Zucearini and doubtless it was one of the garden forms of H. macrophylla having both fertile and sterile flowers. Hydrangea macrocephala Hort. ex Dippel, Handbuch der Laubholz- kunde, 3: 322, 1893. This was a horticultural name published by Dippel as a synonym under his combination Hydrangea opuloides var. a Hor- tensia. It was said to have a very large inflorescence and doubtless refers to a garden form of H. macrophylla having all sterile flowers. Hydrangea mandschurica Hort. ex Dippel, /. c. Hydrangea nigra Hort. ex Dipple, 1. c., 323. Hydrangea mandschurica and H. nigra were horticultural names which Dippel published as synonyms under his combination H. opuloides var. cyanoclada, which he described as having dark purple young shoots and rose-colored, mostly sterile flowers. Hydrangea Sieboldii Hort. ex Dippel, l. c. Another horticultural name, this was published by Dippel under his combination H. opuloides var. Hor- tensia forma macrosepala, which he said had unusually large sterile flowers, and would doubtless be a garden form of H. macrophylla. Section II. CORNIDIA Section Cornidia (Ruiz & Pavén) Engler in Engler & Prantl, Natiirlichen Pflanzen- familien, 3, 2a, 76, 1891. Cornidia Ruiz & Pavon, Flora Peruviana et Chilensis Prodromus, 53, t. 35, 1794, as genus. Sarcostyles Pres] ex Seringe in DeCandolle, Prodromus, 4:15, 1830, as genus. Evergreen climbers, clinging by aerial rootlets, occasionally erect shrubs; leaves coriaceous; bracts few, broadly ovate, at base of inflor- escence, enveloping the unopened inflorescence, leaving noticeable scars on falling; stamens 10; ovary inferior; capsule truncate at apex; seeds elliptical. Mexican, and Central and South American species, except Hy- drangea integrifolia which is Philippine and Formosan. Type species: Hydrangea Preslii Briquet (based on Cornidia umbellata Ruiz & Pavon). 232 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Subsection 1. MoNosEGIA Hydrangea, sect. Cornidia, subsect. Monosegia Briquet, Annuaire Conservatoire et Jardin Botaniques de Genéve, 20:396, 1919. Inflorescences consisting of single terminal cymes; flowers white or pink; sterile flowers present or absent. TYPE SPECIES: Hydrangea Preslu Briquet. 12. Hydrangea Seemannii Riley, Kew Bulletin, 207. 1924. Woody climbers; branchlets ferruginous pubescent with stellate hairs about 0.3 mm. long; leaves oval, 5-20 em. long, 1.5-6 em. wide, about 3 times as long as wide; upper surface glabrous, lower with few stellate hairs; margins entire; petioles 1-5 em. long; inflorescence glabrous; flowers white; sterile flowers usually present; hypanthium 1.5 mm. long; calyx lobes 4 or 5, 0.2-0.3 mm. long; petals 4 or 5, 2 mm. long, 1 mm. wide; stamens 8 or 10, 2.8 mm. long; styles 2 or 3, 1 mm. long during anthesis, clavate; capsule 2.5 mm. long, 3.5 mm. wide, mature styles 1.5 mm. long. DISTRIBUTION: Durango, Mexico, in the Sierra Madre Occidental, at approximate elevations of 2,000-2,600 meters. Type LocALity: Mexico. Durango: ‘in barraneas climbing and rooting on old trees, like Ivy,” at a locality in the Sierra Madre Occidental of Mexico called Rancho de Guadalupe between Mazatlan and Durango, at 6,000 feet. Type collection: Seemann 2142 (holotype, K, isotype, BM). SPECIMENS EXAMINED: MEXICO. Durango: Sierra Madre. Seemann 922 (K); about 35 miles west of El Salto, 2,400-2,500 m., McVaugh 11536 (MICH, CAS); San Luis, 2,600 m., Maysilles 7223 (MICH, CAS). Hydrangea Seemanni is the only species of the genus known to occur in Mexico, where it has been found in the Sierra Madre Occidental be- tween Mazatlan and Durango in the State of Durango. In flower H. Seemann is readily recognized for it is one of two American species in the subsection Monosegia having white flowers. Its fruits, however, are indistinguishable from the pink-flowered H. Oerstedu and H. peruviana, and Seemann’s specimen, a fruiting one with faded sterile flowers, was generally placed in one or the other of these species before H. Seemanniw was created for it. H. Seemann has been known only from the type specimen until the collections of MceVaugh and Maysilles were made. The collection of Maysilles, a flowering one, has established the true relation- ship of this species. Vout. XXIX|] McCLINTOCK: MONOGRAPH OF GENUS HYDRANGEA 233 13. Hydrangea asterolasia Diels, Notizblatt Koniglicher Botanischer Gar- ten und Botanisches Museum Berlin-Dahlem, 15: 370. 1941. Hydrangea epiphytica Morton ex Haworth-Booth in M. Haworth-Booth, The Hy- drangeas, p. 66, 1950. (Published without Latin diagnosis and without desig- nation of a type specimen; however, Skutch 3342 from Heredia, Costa Rica, was the basis of Morton’s herbarium name.) Woody climbers or shrubs; branchlets and inflorescence ferruginous tomentose with lax stellate hairs 0.5—0.8 mm. long; leaves oval, 5-10 em. long, 3-5 em. wide, 2—2.5 times as long as wide; upper surface glabrous; lower surface and petioles tomentose with gray or brown hairs; flowers white; sterile flowers usually present; hypanthium 1.5 mm. long; calyx lobes 4+, 0.5 mm. long; petals 4, 1.5-2 mm. long; stamens 8, rarely 10, 3-4 mm. long; styles 2, 1-1.5 mm. long during anthesis, clavate; capsule 2 mm. =] Map 9. Distribution in Costa Rica and Panama of A Hydrangea Oerstedii M@ Hydrangea asterolasia @ Hydrangea peruviana Pt we : F a e a iF Ay ; og . : 7 v2 -_ ‘ = 7 4 ¢ RS ets F ay ot Pa ee : : - 4 2 : aes it on , , a. te ery se aerryeere. - 7 7a ae A axe ty | “oes vs ale, 5 iy 307 hiterature- Cited aeeerst ke eae, oF ea. chet eis Oe a ee 309 INTRODUCTION Isla Guadalupe, an oceanic island rising from deep water off Baja Cali- fornia, about 225 nautical miles south and slightly west of San Diego, and separated by 133 miles from the nearest other land (Isla San Gerénimo), is proving to be characterized by a high incidence of endemism in its littoral marine as well as in its terrestrial biota. The marine endemiecs include a considerable number of littoral fishes. A few of these, including two gobi- esocids discovered by the senior author and recently described by Briggs (1955: 79-80 and 100-101, figs. 5, 95, and 100), have been proposed as full species. Others remain to be named, either as species or subspecies (the problem is currently under study by the senior author and Erie Godwin Silas). Of the apparently endemic species perhaps the most strikingly distinctive is the butterflyfish described in this paper. Except as noted, the systematic methods, especially of counting and measuring, are those proposed by Hubbs and Lagler (1952). Angles were measured as proposed by Hubbs (1946). We are indebted to a number of colleagues for assistance in the examina- tion of material, for data, and for suggestions in the presentation. In par- ticular, we acknowledge such help from Leonard P. Schultz of the United States National Museum, who verified our conclusions regarding the value of the lateral-line character and the placement of the species nigrirostris, who made material in the National Museum freely available, and who had two photographs prepared. We are similarly grateful to Loren P. Woods of the Chicago Natural History Museum, who, like Schultz, provided much information on the lateral-line characteristics of many species and offered valued suggestions regarding the taxonomy of the group, which he has extensively studied. Woods also loaned us specimens of Chaetodon aya. We express appreciation also to Bovd W. Walker, our colleague at the Uni- Vou. XXIX] HUBBS & RECHNITZER: A NEW CHAETODON 275 versity of California at Los Angeles, who has made available extensive material of the previously known eastern Pacifie chaetodontines and has otherwise been of help and encouragement in the preparation of the paper. Robert L. Wisner made the X-ray photograph reproduced as plate I, figure 2. Max Poll of the Koninklijk Museum van Belgisch-Kongo has allowed the reproduction of the type figure of Chaetodon marcellae. W. I. Follett and Lillian Dempster of the California Academy of Sciences have assisted generously in the processing of the manuscript. A NEW SPECIES FROM GUADALUPE ISLAND Chaetodon falcifer Hubbs and Rechnitzer, new species. (Plate I.) TYPE SPECIMEN A large adult, 138 mm. in standard length, collected at a depth of 100 feet on the east shore of Guadalupe Island, off Baja California, México, on November 16, 1954; Cat. No. 20734, Department of Fishes, California Acad- emy of Sciences. CoLLECTING NOTES AND HABITAT DATA The only known specimen of this species (pl. I, fig. 1) was speared by Rech- nitzer during a deep-poisoning operation on November 16, 1954, supplemented with detonation of Primacord on November 17, in a cove on the generally calm eastern shore of the voleanic island, about 0.5 mile north of Morro Sur (South Bluff), at 28° 53’ 24” N. Lat., 118° 15’ 00” W. Long. (as measured on Hydro- eraphie Office Chart 1688 (1956) ; on the older chart, 1681, the island was mis- placed 1.5 mile to westward and 0.6 mile to southward). Rechnitzer swam down with aqualung equipment to try to collect a bright orange-yellow fish (a pomacentrid?) that he had seen while applying the derris-root paste at a depth of about 20 feet. Failing to relocate that specimen in the clouded water, he went lower, to the rocky bottom at a depth of about 100 feet, where, about a foot off the bottom, under a slight ledge at the base of the vertical underwater cliff, he saw, in the extremely clear water, and speared, the magnificent butterflyfish we are describing. Where the fish was obtained, plants, other than encrusting ecoralline algae, were few and seattered. In shallower water there was much attached Sargassum and other algae and some eelgrass (Zostera). The temperature was recorded by one of the divers as 65° F.. from top to bottom. In addition to rock there was some sand and gravel on the level bottom, which was about 100 feet deep. The vertical rock wall was penetrated by deep recesses and vertical crevices, which provided excellent cover for reef fishes. Bearing in mind the favorable habitats at this station and the large ~] a CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH Ser. VoL. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 277 party of collectors (thirteen, including six with aqualungs), it is not sur- prising that a considerable proportion of the depauperate littoral fish fauna of Guadalupe was taken with the Chaetodon. The thirty-three species listed below were included in the collection and are therefore to be listed as ap- proximate associates of the butterflyfish. Species only observed are marked with an asterisk. Endemies are marked with a dagger. Sternoptychidae Vinciguerria lucetiae Garman Exocoetidae *Oypselurus californicus (Cooper) Apogonidae Apogon guadalupensis+ (Osburn and Nichols) Serranidae Paralabrax clathratus (Girard) Atherinidae Atherinopsis californiensis subsp.+ Atherinops affinis guadalupae+ Hubbs Carangidae *Seriola dorsalis (Gill) Trachurus symmetricus (Ayres) Branchiostegidae Caulolatilus princeps anomalus (Cooper ) Embiotocidae Embiotoca sp., possibly differenti- ated Damalichthys sp.+ Brachyistius aletes+ (Tarp) Pomacentridae Chromis punctipinnis (Cooper), somewhat differentiated Azurina hirundo+ (Jordan and Mc- Gregor ) Hypsypops sp.} Labridae Bodianus diplotaenia (Gill) Pimelometopon pulchrum (Ayres) Girellidae Girella nigricans (Ayres) Scorpididae Medialuna californiensis (Stein- dachner ) Scorpaenidae Scorpaena guttata guadalupae Fow- ler (occurs also in Gulf of Cali- fornia ) Scorpaenodes xryris (Jordan and Gilbert), possibly differentiated Cottidae Artedius creaseri (Hubbs), possibly ditferentiated Gobiidae Lythrypnus zebra (Gilbert), pos- sibly differentiated Lythrypnus dalli (Gilbert), possibly differentiated Clinidae Alloclinus holderi (Lauderbach) Heterostichus rostratus guadalupen- sis; Hubbs Gibbonsia elegans erroli+ Hubbs Gibbonsia norae Hubbs, also on San Benito Islands Tripterygiidae Enneapterygius sp., not endemic Brotulidae Eutyr diagrammus Heller and Snod- grass, also on Galapagos Islands Halichoeres semicinctus (Ayres) Balistidae Oryjulis californica (Giinther), pos- Xanthichthys lineopunctatus (Hol- sibly differentiated lard) PLATE I Figure 1. Holotype of Chaetodon falcifer, 138 mm. in standard length, from Guadalupe Island, México. Photo by Scripps Institution. Figure 2. X-ray photograph, by Robert L. Wisner, of holotype of Chaetodon falcifer. Print by Seripps Institution. 278 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. The collecting at this station (SIO 54-219) was a joint operation of two parties, largely from the Scripps Institution of Oceanography, led, respec- tively, by the junior author on the Institution’s Research Ship Paolina T, and by the senior author on the Research Ship Orca, of the J. W. Sefton Foundation. Thanks are due to all members of the dual expedition, espe- cially to Joseph W. Sefton, Jr., who, often and graciously, made the Orca available for biological explorations. DISTRIBUTION It is possible, indeed highly probable, that this apparently deep-water Chaetodon is much commoner about Guadalupe Island than the collecting of only one specimen would indicate. However, no other specimen has been seen, despite the fact that, of late, the island has been subjected not only to rather intensive collecting but also to a considerable amount of aqualung diving by several experienced and observant divers, including Conrad Lim- baugh. During the expedition of November, 1954, six divers with self-con- tained underwater breathing apparatus were under water much of the time during seven consecutive days. The center of abundance of the species may well be deeper than 100 feet (as seems to be true of the related Atlantic species) ; or, 1t may occur chiefly on the western shore of the island, where the surf is usually strong and where collecting below tidal limits has been very limited. It is also possible, though we think rather improbable, that the species occurs farther south, either along the mainland or on the Revillagigedo Islands. It has not been included among the chaetodontids taken during the extensive collecting in those areas in recent years, particularly by parties working under or for Boyd W. Walker of the University of California at Los Angeles and John E. Fitch of the California Department of Fish and Game. A considerable amount of more or less deep-water poisoning has been included in those operations. For these reasons, Chaetodon falcifer may be listed, provisionally at least, among the endemic fishes of Guadalupe Island. The interpretation of C. falcifer as a relative of the deep-water Atlantic species, C. aya and C. marcellae, weakens but hardly destroys the indication of endemicity. Even though endemic, Chaetodon falcifer is to be included among the tropical elements in the intermixed warm-temperate and tropical fish fauna of Isla Guadalupe. In the eastern Pacific it probably represents the northern limit in the present distribution of the tropical family Chaetodontidae. The qualification “present” is used because, about one hundred years ago, when the fauna of southern California was distinctly more tropical than at present (Hubbs, 1948), Chaetodon humeralis Giinther seems to have occurred at VoL. XXIX] HUBBS & RECHNITZER: A NEW CHAETODON 279 San Diego. We are informed by Leonard P. Schultz that two specimens of that species in the United States National Museum are recorded as having been collected by Lt. Trowbridge at San Diego, California. The low cata- log number (3170) indicates that the specimens were in fact entered in the collection nearly 100 years ago. Lt. W. P. Trowbridge was one of the most effective of the West Coast collectors on the Pacific Railroad Surveys. Somehow the species escaped inclusion in Girard’s reports on the fishes collected by these surveys. Presumably the specimens came to light after the bulk of the collections had been studied and cataloged, for the number is higher than those recorded for the species reported by Girard. In this connection it may be noted that Girard (1858: 338) referred to other speci- mens from San Diego that had become “‘mislaid in the moving of the Smith- sonian collections from one end of the building to another a few months since.” Some slight doubt regarding the validity of the San Diego record can not be dispelled, but we believe that C. humeralis is to be added to the list of tropical fishes that occurred at San Diego during the warm period a century ago. DISTINCTIONS Chaetodon falcifer is not closely related to any of the three chaetodon- tines already known from the eastern Pacific. It does not belong in the same subgenus as C. humeralis Giinther (1860: 19-20; 1869: 419, pl. 65, fig. 3), which is the only one of the three that we retain in the genus Chaeto- don. It agrees with that species, as well as with the two others, in the high spinous dorsal with deeply incised interspinal membranes, but it has a much sharper and more produced snout and very much smaller scales, and the scale rows on the middle of the trunk are subhorizontal and those above the lateral line are parallel with the lateral line (instead of being strongly oblique in each area). Furthermore, the coloration is strikingly different, and there are numerous trenchant distinctions in the bony and sealy strue- tures of the head, and in the squamation of the caudal fin (see pp. 295-296). The Guadalupe species differs even more significantly, we believe, from the other eastern Pacific form currently referred to Chaetodon, namely C. nigrirostris (Gill). In that species, as is indicated later (pp. 296-299), the lateral line continues to the caudal base: a character that seems to eall for its placement in another genus. Provisionally, we refer it to the genus Heniochus. Chaetodon falcifer and Heniochus nigrirostris differ in many other respects, as in coloration, in the squamation of the head, in the char- acters of the gape and lips, ete. The differences can be appreciated by com- parin® the following deseription of C. falcifer with the descriptive notes on H. ngrirostris (pp. 296-297). 280 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH Ser. When compared with Forcipiger longirostris (Broussonet), C. falcifer is seen to differ not only in the much less extremely produced beak, but also in the lateral-line character, in which Forcipiger and Heniochus agree, and in many features of coloration. Inasmuch as some Indo-Pacific fishes, including one of the three eastern Pacific chaetodontines, have hurdled the broad barrier between the central Pacific islands and the New World (Ekman, 1953; Myers, 1941, and others) to become established in the Panamaiec fauna, C. falcifer has been compared with descriptions and figures of the multitudinous Indo-Pacific species of the Chaetodontinae. In this comparison numerous treatises have been con- sulted, including Bloch (1796), Cuvier (in Cuvier and Valenciennes, 1831), Giinther (1860), Bleeker (1876a—1878), Jordan and Fowler (1902), Jordan and Evermann (1905), Jordan and Seale (1906), Kendall and Goldsborough (1911), Ahl (1923), Herre and Montalban (1927), Fowler (1928, and supplements), Fowler and Bean (1929), Weber and de Beaufort (1936), and Munro (1955). Also checked were species listed in the Zoological Record as having been published since the appearance of Ahl’s monograph in 1923. Nothing corresponding closely with C. falcifer has come to light. It is coneluded, therefore, that this species is not a stray from the present Indo-Pacific fauna. On comparing C. falcifer with the Atlantic species of the genus, we find marked differences, but many characters shared with C. aya Jordan of the western Atlantic and C. marcellae Poll of the eastern tropical Atlantie (see pp. 300-304). There are consistent differences in the color marks. In the Pacific species the preocular part of the first band runs almost hori- zontally along the upper part of the side of the snout, rather than obliquely across the front part of the cheek just behind the mouth, and the posterior mark has a seythelike form with an anterior arm extending down to the opercular region, rather than being confined to a simple bar running obliquely downward and backward across the body from near the middle of the spinous dorsal (in C. aya), or subvertically downward from the end of the first dorsal (in C. marcellae). Furthermore, the scales are much finer in C. falcifer than in the Atlantic species, and there are many other, minor differences. A careful comparison of the type of C. falcifer with the 17 available specimens of C. aya discloses the numerous differences outlined in table 1. Chaetodon falcifer apparently differs from C. marcellae in most of the characters that distinguish it from C. aya. 281 A NEW CHAETODON HUBBS & RECHNITZER: Vor. XXIX] *po10[d UOl}B1S JSOULUAIYILIOU 9} YIM SoploULoOdD A|[BIVUGSSe yng ‘UWOISN[DUL LO] 9}¥] 00} POATIIAI SBM (69EF9 ‘IN’ H'N’'O) Uoeultoeds 90 IOJ UOT}EIS PLOdDAL ey, ‘seteds poel[[e pure waf1a)vf UOpojJanYy) AOJ SUOTJBIS pLOdII UMOUY ‘LT Ae] q JIMAaWY ALNOS — _3V11394vW NOGOL3VHD 4 wadld1V4 e@ NOGOL3IVHD VYVIIYAIWY HLYON 282 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH Serr. Table 1. Comparison of Chaetodon falcifer with C. aya Chaetodon falcifer Chaetodon aya Dorsal soft-rays 21 17 to 19 Anal soft-rays 16 14 to 16, usually 15 Pectoral rays 15 13 to 15, usually 14 Scale rows, straight line? 51 to 53 32 to 38 Seales in first row above lateral line? 64 to 66 27 to 36 Pores in lateral line® 44 to 46 23 to 32 Seales from dorsal fin to lateral line‘ 5 9 to 12 Scales from anal fin to lateral line‘ 26 17 to 21 Scales around caudal peduncle? 30 20 to 26 Lateral line ends: Close to end of soft Below front part of dorsal soft dorsal® Main marking on body’ Scythe-shaped, with Oblique band, extending extension onto opercles onto spinous dorsal Along side of snout To just back of upper jaw With much black, and Almost clear, except on a definite light border extension of bars; no definite light edge Band below eye extending: Dorsal and anal fins Upper and lower borders of caudal fin Largely blackish Not blackened Angle, body axis and margin of soft dorsal® 72° 75° to 87° Maximum known size 138 mm. 98.5 mm. RELATIONSHIPS We have not been able to allocate C. falcifer with complete assurance in any of the named subdivisions of Chaetodon, as these have been charac- terized in various published revisions (subsequent authors have based their eroupings, though not their nomenclature, largely on the subgenerie divi- sions proposed or accepted by Bleeker from 1876 to 1878). When checked with Ahl’s monographie revision of 1923, falcifer is seen 1 Rows crossing a straight line from scapular process behind upper end of gill opening to caudal flexure; for this and for all other paired series, both sides were counted on all specimens. 2 Including all scales that impinge above on a lateral-line scale. ® Including occasional unpored scales definitely in the lateral-line series. 4 From edge of sheath at side of first spine in an oblique row to but not including the lateral line. ® Scales in not quite overlapping position around narrowest part of peduncle. 6 Ending rarely, on one side only, on vertical passing slightly before visible front of soft dorsal or a little behind middle of soft-dorsal base. 7 When considering colorational characters. compare plate I, figure 1 with plate II, figure 1 and plate III, figure 1. 8 Angle between body axis (from front of snout to middle of caudal base) and a symmetrical tangent to the rounded margin of the soft dorsal fin. VoL. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 2 oO eM) to agree essentially with subgenus Chaetodon, and not with the subgenera Gonochaetodon and Megaprotodon, which Weber and de Beaufort (1936) later recognized as genera. When compared with the three divisions of “Gruppe” rank into which Ahl split the subgenus Chaetodon, respectively with two, four, and three further subdivisions ot “Sektion” rank, falcifer is excluded from group | by reason of the body form and from group 3 by having the median trunk seales little enlarged, with well-rounded margins and with the subhorizontal alignment retained; also in having more than 45 scale rows crossing the median line of the body. This leaves “Gruppe II,” from sections D to F of which falcifer is excluded by the sub- horizontal scale rows. From “Sektion C, Chaetodon s. str. = Tetragonoptrus Bleeker,” as briefly diagnosed by Ahl (p. 80), falcifer is not obviously ex- cluded, but it has smaller scales than any of the species as counted by Ahl, except the very different “Chaetodon Sanctae Helenae Giinther.” In fact, the scales are rather too high for his generic diagnosis (p. 7): Ahl characterized the genus Chaetodon as having about 30 to 50 scales in “L. lat.” (defined on p. 6 as the number ‘‘vom oberen Ansatz des Kiemendeckels” to the caudal flexure). Ahl thus contrasted Chaetodon with Hemitaurich- thys, Microcanthus, and Vinculum, each of which was indicated as having 60 or more (60-90) scale rows. Those three genera all have the lateral line of the primitive Heniochus type, and one of them, Microcanthus, has been shown by Fraser-Brunner (1946) to have been wrongly classed in Chaetodontidae. Chaetodon falcifer differs more trenchantly from any of the species placed by Ahl in his section Chaetodon in having a much longer, more pro- duced, more pointed snout, in this respect definitely approaching Forcipiger and Chelmon. It agrees fully with none of the subdivisions of Chaetodon, including the segregated genera Gonochaetodon, Megaprotodon, and Ani- sochaetodon, as diagnosed by Weber and de Beaufort (1936). Further studies may eall for its recognition as a member of a new subgenus, or pos- sibly genus (presumably to include also the Atlantic species, discussed below). Several characters that Chaetodon falcifer shares with Heniochus nigri- rostris and Forcipiger longirostris (Broussonet), which is regarded as in- eluding F. flavissimus Jordan and McGregor, suggest the remote possibility of a relationship with one or both of those species; close enough, if verified, to run counter to the classification we have adopted. Chaetodon falcifer agrees with H. nigrirostris in the size of the seales, in which respect C. falcifer contrasts with almost all species currently left in Chaetodon. It also agrees with that species, though not so exclusively, in the orientation of the seale rows. In the form of the spinous dorsal the 284 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH Serr. two species are similar, having strong and high spines with very deeply incised membranes; but in C. falcifer the third and fourth spines are sub- equal, whereas in H. nigrirostris the fourth spine is definitely the longest, as it is (to a greater degree) in the other species referred to Heniochus (furthermore, in the apparently closely related Atlantic species, C. aya, the third spine is definitely the highest). There are also some resemblances (as well as sharp differences) in the color pattern, for both species lack the suborbital bar characteristic of most species of Chaetodon and have the nuchal bar continued forward on the snout (though with interruptions in H. ngrirostris). The posterior dark crescent in H. nigrirostris suggests the posterior part of the seythelike mark of C. falcifer, though it is farther back and farther up. Chaetodon falcifer approaches Forcipiger longirostris in the sharpness and production of the snout and in the reduced size of the scales. In both species the dorsal spines are high and the interspinal membranes are deeply incised. In both, the nuchal band is continued through the eve to and along the upper part of the snout. In some of these characters all three species under discussion show resemblances. At present, however, we interpret each of these resemblances as the result of convergent evolution. Numerous differences between all the forms may be noted by comparing the following description of C. falcifer with the descriptive notes of the three other Pacific species (pp. 295-300). As already suggested (p. 00), C. falcifer seems to find its closest rela- tives in two Atlantic species, C. aya of the western Atlantic and C. mar- cellae of west Africa, which are treated in more detail on pages 300-304. The distinctive resemblances (compare figures) include the sharpness and pro- duction of the muzzle (a character that is carried to a farther extreme in Prognathodes aculeatus of the West Indian fauna); the very high, strong, and greatly excised dorsal and anal spines; the rather regular and little modified alignment of the scale rows; the rectangular form of the body behind the head; the more or less forward dislocation of the subocular bar (carried to an extreme in C. falcifer); a dark bar along the top of the head anteriorly; the presence and intensity of the blackish bar from eye to front of dorsal; the extension of the posterior bar to an abrupt end on the seale- covered base of the anal fin, ete. These resemblances seem to outweigh the sharp differences between C. falcifer and C. aya that are outlined in table 1, and that also apply largely to C. marcellae. The close relationship between C. falcifer, C. aya, and C. marcellae is further suggested by their distinctively deep-water habitats. The type of C. falcifer, as indicated earlier, was collected at a depth of 100 feet. All specimens of C. aya and C. marcellae, so far as recorded, were taken at Vou. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 285 even greater depths. The type of C. aya was from the stomach of a red snapper from Snapper Banks off Pensacola, Florida, where the depths fished at that time were reported by the collector, in his survey of the fish- eries grounds (Stearns, 1887), to be 20 to 50 fathoms. Another specimen was trawled off North Carolina (Nichols and Firth, 1939: 87) at a reported position where the coast chart (1110) indicates the depth as approximately 150 fathoms (considerable error either in distance or direction would not fix the trawling spot at a depth of less than 100 fathoms)! Hildebrand (in Longley and Hildebrand, 1941: 150-151) reported the capture of seven specimens near Tortugas, Florida, at two stations, at depths of 39 and 40 fathoms, respectively, and added, “apparently not entering shallow shore waters.”’ The additional known specimens of C. aya (see pp. 302-303) have been taken at depths recorded as 24, 25 to 30, 25 to 75, 37, 40 to 45, 62, and 75 fathoms. DIAGNOSIS Outstanding characters of Chaetodon falcifer are the very long and sharply produced snout, in which respect it definitely approaches For- ciprger; the rather small scales, intermediate in size between those on most species of Chaetodon and those typical of the species of Heniochus and re- lated genera; the lack of a suborbital bar and the extension of the nuchal band, as in Forcipiger, through the eye to and along the snout; and the peculiar seythe-shaped blackish mark on the body. Less distinetive, but serving to separate the species from many others in Chaetodon (and segre- gated genera) are the dorsal and anal fin formulas, respectively XIII, 21 and III, 16; the high and strong dorsal and anal spines, with deeply incised membranes; the alignment of the scale rows, which are subhorizontal below the lateral line and parallel with that line above it; the rounded margins and only moderate enlargement of the median trunk seales; the rather rectangu- lar body form; the low mouth; and the alignment of the inner jaw teeth ante- riorly in twelve rather even cross rows. DESCRIPTION OF FINS Dorsau FIN. The dorsal rays number XIII, 21. The base of the spinous dorsal measures 1.6 in the soft-dorsal base. The dorsal spines are very strong and high anteriorly. The subequal third and fourth spines are the highest. Each, as measured from the edge of the scaly sheath, is as long as the soft- dorsal base. The interspinal membranes are deeply incised, very deeply anteriorly. The least width of the membrane at the front of the two longest spines is less than one-sixth the height of the spine. The soft dorsal has a 286 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. weakly rounded margin, with the median part nearly straight and forming an angle of 72° between its tangent and the axis of the body; the margin along the first three or four rays is nearly horizontal; that along the last few rays is rounded and is directed downward and slightly forward. None of the dorsal soft-rays are produced. ANAL FIN. The rays number III, 16. The spines are very strong. The second is very notably strengthened and considerably elongated, though when depressed it just reaches the tip of the third spine. The free front edge of the second spine is as long as the base of the soft anal. The inter- spinal membranes are very deeply incised. The membrane at the front of the second spine is one-sixth the length of the front edge of that spine. The margin of the soft anal is nearly straight and nearly vertical medially, and is moderately curved forward along the anteriormost and posterior- most rays. OTHER FINS. The caudal has a weakly convex border, with the upper angle rather pointed, though not produced, and the lower angle more broadly rounded. The very slightly faleate pectoral, which reaches to above the anus, is about one eye’s length shorter than the head. The first pelvic soft-ray ends in a free filament, which is more than half as long as the orbit and reaches to between the genital papilla and the origin of the anal fin. The length of the pelvie equals that of snout plus orbit. Principal eaudal rays, 9 + 8 = 17; pectoral rays 15 on each side, or, more precisely, distinguishing unbranched and branched rays respectively on left and right sides an, 12) ne a. SCALES AND LATERAL LINE SCALE NUMBER AND SIZE. The seales are relatively small and numerous. Along a straight line from the upper end of the gill opening to the caudal flexure we enumerate 51—53 transverse rows (counting both sides) behind the scapular process, and 4 rows on the process; there are 3 additional rows of seales, smooth and well embedded, on the concealed part of the shoulder girdle. Along the first row above the lateral line, beginning behind the scapular process and enumerating all scales that definitely impinge above on a lateral-line scale, we count 64—66 seales. From the edge of the scaly sheath beside the first spine, in an oblique row to but not including the lateral line, we count on each side 15 seales below the dorsal and 26 above the anal. There are about 30 rows, not including seales that overlap on the exposed field, around the narrowest part of the caudal peduncle. The scales become markedly reduced in size on the thick sealy area over the basal part of the soft-dorsal fin. The scales also become very small on the VoL. XXIX] HUBBS & RECHNITZER: A NEW CHAETODON 287 nape, occiput, opercles, and muzzle. They are somewhat but not notably enlarged on the side of the trunk forward to the pectoral fins, but these larger scales grade evenly into the others. SCALE FORM, SERIATION, AND DISPOSITION. The scales are regular in out- line, with evenly rounded margins, even where moderately enlarged. Below the lateral line the seale rows are approximately horizontal, with a slight upward flexure posteriorly that increases to about 18° on the fleshy area over the anal base. Some rows are arched upward a little where the scales are most enlarged. Toward the pelvic fins the rows become irregular. Above the lateral line the rows are essentially parallel with this line and are similarly curved, on the body proper, but become irregular, and peripher- ally assume an oblique orientation, on the thick sealy fin bases, especially on the second dorsal. Where, on and about the head, the scales become greatly reduced in size, the rows become very irregular or hardly apparent. Very small rough scales are developed on the posterior part of the premaxillary, especially in files along the fine rugose ridges. A narrow band of etenoid scales crosses the top of the premaxillaries just in front of the rostral groove (this char- acter will likely prove variable, as it is in C. aya). The exposed part of the maxillary, behind the anterior grooved area, is evenly covered with small ctenoid seales. Fine ectenoid seales cover the mandibles forward to a little beyond the end of the gape. There is only a small sealeless area around the anterior nostril and below the posterior nostril. This sealeless strip extends backward to the orbit below the anterior prolongation of the upper bony orbital rim. A very narrow and irregular sealeless groove, seemingly distinctive among the American Pacific species of the subfamily, but which will probably prove variable, as it is in C. aya, extends forward from in front of the anterior nostril to the row of seales bordering the rostral edge. A small triangular sealeless area (which will probably prove variable, as it is in other species) extends upward and backward from the upper bony orbital rim. At the apex, the width of this sealeless area is only about one-fifth that of the orbit. The fleshy upper orbital rim, largely concealed by the bony rim, bears some scales, mostly small and smooth (probably also a variable character, or one affected by loss of scales). The sealeless margin of the soft dorsal is about one-third as wide as the orbit; that of the soft anal, about one-fourth the orbit. In each fin the outer part of the interradial membranes is more or less sealeless for an addi- tional distance about equal to the width of the totally sealeless margin. Anteriorly, these sealeless grooves are more extensive. Along the thickened base of each alternate dorsal spine and along one side of the third anal spine an acute triangle of scales runs well out on the spine. Except on a basal band about two-thirds as wide as the orbit, and near the upper and 288 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. lower margins, the caudal fin is very weakly scaled. The scales on the branching rays are very small, delicate, and difficult to see. Beyond the scaly base single or double files of scales extend only a very short distance outward on each interradial membrane, scarcely at all on the membranes between the innermost rays. (It is possible that etenoid scales may have covered more of the caudal in life.) LATERAL LINE. The lateral line forms an almost even are backward to its end about five tiny scale rows before and above the end of the dorsal base. The pores behind the scapular process number 44-46. The small lateral-line scales are more isolated from one another than in C. aya. On the midsides in advance of the vertical from the end of the lateral line, on the right side only, are two pored scales in series (not included in the count). Bopy AND HEAD CHARACTERS Bopy Form. Behind the head, the body, including the fleshy sealed fin bases, is virtually square. The vertical depth from the edge of the dorsal scale sheath to near the pelvie insertion steps 1.75 in the standard length. The edge of the thick sealy covering of the dorsal fin is approximately straight and horizontal from the second dorsal spine to the extreme front part of the soft dorsal. The ventral contour is similarly almost straight and horizontal from a little in front of the pelvic insertion to the extreme anterior part of the soft anal. From the dorsal origin to the concavity of the snout the profile is almost straight (very slightly sigmoid) and forms an angle of 52° with the body axis and an angle of 83° with the chord of the anteroventral contour behind the muzzle. The scaly bases of the soft- dorsal and soft-anal fins appear very steep in external view, though the actual structural bases as measured on the X-ray photograph (plate I, figure 2) describe an angle of only 81°. SNOUT AND MUZZLE. The snout is notably produced, narrowed, and sharpened. Its dorsal profile forms an angle of 150° with the profile between the snout and the dorsal fin. The angle of the muzzle is about 40° in lateral view and 29° in dorsal aspect (each measurement excluding the abruptly rounded tip.) The greatest width in front of the eye steps 1.8 in the length of the snout. Since the snout is a little longer than the postorbital, the center of the eye is slightly behind the middle of the leneth of the head, which, including the opercular membrane, steps 2.8 in the standard length. MoutH ANpb jaws. The front of the low and subhorizontal gape lies definitely below the lower border of the orbit. The lower border of the upper jaw forms an angle of 21° with the body axis. The gape is not quite half the leneth of the upper jaw, which enters the head length 3.5 times. The anterior parts of the jaws are expanded forward to accentuate the beaklike form of the snout. The midline length of the upper lip is actually Vou. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 289 more than half the length of the orbit. Above and behind a deep fissure the lips are sculptured by parallel grooves and more or less crenate ridges. The fissure is about three-fourths as long as the orbit. It starts, in side view, near the middle of the anterior premaxillary expansion and extends to approximately the middle of the length of the posterior maxillary dila- tion, near its lower border. There is no vertical groove on the maxillary. The margin of the upper lip is somewhat pendant, especially on the sides opposite a shght anterior lobation of the lower lip. The front of the gape is narrowly and evenly rounded transversely, with no marked irregularity on either lip. TEETH. The teeth of the jaws extend to the end of the gape. Those of the outer row are slightly enlarged. The very fine inner teeth are defi- nitely aligned in each jaw in twelve nearly straight transverse rows across the semispherical tip of the beak. Posteriorly the inner teeth are less regu- larly arranged and form a narrower band. The midline length of each anterior tooth patch measures 3.5 in the orbit. The vomer is edentulous. GILL RAKERS. The rakers, including rudiments, number 5 + 11 on the outer arch. Those of the upper limb are soft. The longest, near the middle of the lower limb, barely reach the second raker below. Bony MARGINS. The preorbital border is finely denticulate, especially at the squarish anterolateral corner. The preopercular denticulations, mostly very fine, are somewhat enlarged and directed backward at the rounded angle, above which the nearly vertical margin is rather strongly coneave and below which the somewhat oblique edge is weakly concave. The exposed border of the scapular process is also denticulate. The upper border of the orbit is roughened by several series of minute denticulations. In the sealeless area, extending upward and backward from the upper orbital rim, the fine spinulation continues for a short distance and then becomes transformed into larger points and very short ridges, rendering the whole exposed area rough (perhaps more so than in other specimens). The bony orbital rim is scareely expanded in the prefrontal region. It is extended downward and forward about halfway to the posterior nostril, departing here from the margin of the orbit and lying above the sealeless area behind the nostrils (this character will probably vary widely, as it does in other species). The bony orbital margin is extended down the posterior edge, to continue below the orbit and up to near the middle of the front edge. Around the lower half of the orbit the more or less trenchant bony rim is armed in the type specimen with a row of fine serrulations, in places backed by other, minute points. OSTEOLOGICAL F’'EATURES VERTEBRAE AND HYPURALS. Many of the bony structures can be made 290 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. out on an X-ray photograph (plate I, figure 2). The vertebrae number 10 + 14 (including the hypural plate). The first two are reduced in size. The caudal rays impinge on eleven hypural elements, six above and five below, of which 3 + 3 comprise the hypurals proper, arising from the lower edge of the upturned axis. The hypurals support all seventeen prin- cipal caudal rays except the lowermost two, which arise, respectively, just above and on the more posterior of the two hypural-like interhemals of the twenty-third vertebra. On the six hypurals the principal rays originate according to the following formula, counting downward: 3 + 4 + 2— 9; 1+ 4+ 1=6. The four upper procurrent rays arise from the two loose epurals, and a minute blob of ray tissue hes beyond the tip of the neural element of the last vertebra. The three lower procurrent rays arise from the two hemal elements connected with the last normal centrum. Only the posteriormost procurrent ray, above and below, is segmented. VERTEBRAL SPINES AND INTERSPINALS. The interneurals of the last two dorsal spines arise on either side of the definitely longer neural spine of the twelfth vertebra. This neural spine is sagittally expanded toward the base, thus contrasting with the spines that follow. From the twelfth to the eighth the neural spines remain nearly at right angles to the vertebral column, in- crease in thickness, and become shorter, finally to reach only halfway to the base of the dorsal spines. From the seventh to the third the neural spines become oblique and still shorter. The first two become more erect and the first is the smallest. The interneurals increase in strength forward from that of the last dorsal spine to that of the second spine, which inter- neural is directed slightly forward to fit between the first two neural spines close to the vertebrae. The interneural of the first dorsal spine comprises a flat shaft that impinges on the following interneural and that bears, at its upper end, a strong thornlike process directed forward and downward. The two rodlike interneurals that arise on either side of the tip of the first neural spine have, at the top, nearly fused, forward-directed, thornlike processes, which, together with the larger process just mentioned, form a bridge, hidden beneath the sharp anterodorsal rim of the body, between the first dorsal spine and the very strong, high, and sharply elevated supraoccipital spine. The hemal spine of the first caudal vertebra is considerably expanded where it lies behind the very strong spinelike interhemal of the second anal spine. That interhemal reaches slightly more than halfway to the central line of the vertebral column. SKULL. Above the orbital rim and the base of the first vertebra, the skull in side view forms approximately an equilateral triangle, with the apex, at the base of the supraoccipital spine, located above the isthmus a distance equal to that from the tip of the snout to the posterior margin of VoL. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 291 the shoulder girdle. In side view the premaxillaries and dentaries are very slender. The least dorsoventral dimension of the premaxillaries, near the rear of the anterior dental expansion, is only 0.1 the antorbital leneth of the skull. COLORATION (Plate I, figure 1) SCYTHE-SHAPED MARK. The outstanding colorational feature of this spe- cies is the scythelike mark on the sides of the head and body. In life this mark was deep brown-purple, almost jet-black. After over a year in 40% isopropyl aleonol, following initial preservation in formalin, this color, like many of the others, is little changed. The short “handle” of the “scythe” begins abruptly at the upper-posterior part of the interopercle, beneath and behind the preopercular angle. The “handle” continues upward and slightly backward to cover the subopercles and the opercle behind the upper bony process. There are dusky dapplings on the adjoining branchiostegal membrane. Continuing its steep course, at an angle of 62° with the axis of the body, the mark crosses the opercular membrane and attains a position astride the lateral line. Along the middle third of its anterior section, where it is slightly arched backward, the mark is about as wide as the orbit. The front edge of the band is weakly arched forward dorsally and extends to a rounded angle, somewhat greater than a right angle. The angle of the axis is about 80°. The apex is separated from the edge of the sealy sheath near the front of the fifth dorsal spine by an interval about two-thirds as wide as the orbit. The lower-posterior border of the anterior section of the mark follows throughout most of its length a nearly straight course, a little steeper than the anterodorsal contour of the body. Dorsally the ventral border arches rather abruptly backward, to reach its apex below the posterior exposed base of the fifth dorsal spine, in the third scale row below the lateral line. From the apex the lower border is gently curved to the acute end of the “seythe” on the sixth anal soft-ray. The upper border continues backward and slightly downward, nearly straight, to descend below the lateral line at a point nearly in line with the last spine of the normally expanded dorsal fin; that is, at a point about midway between the apex of the mark and the end of the dorsal base. From this point the upper edge of the “seythe” arches backward and then downward to the lower-posterior end. On the fleshy base of the anal fin the posterior edge of the mark is nearly straight and nearly vertical. The mark is separated from the end of the dorsal base by a distance two-thirds the leneth of the orbit; from the caudal flexure, by one orbital length; and from the end of 292 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. the anal base, by about one-fourth the orbit. The width of the band, meas- ured either downward and backward from the anterior apex, or at the oreatest width posteriorly, steps 2.4 in the head length. The least width between those points, not far behind the apex, is 3.3 in the head. BAND THROUGH EYE, AND ADJACENT CoLor. Another prominent and con- spicuous mark is the band, rather rich-brown in life and sooty blackish in preservative, that extends from the base of the first two or three dorsal spines downward and slightly forward to the upper rim of the orbit. With decreased intensity, though still conspicuous, it cuts obliquely across the eye and then continues, much less steeply and with a weak downward curvature, well above the mouth, toward but not quite to the rostral fold; except near the eye, the lower border is horizontal. The narrow light area in front of the band was somewhat yellow in life. Above and behind the eye there was considerable yellow both before and behind the band. On the head above the eye the mark is blackest, with a further intensification along either edge. Farther upward and backward the anterior edge remains sharp, along the narrow light V on the predorsal contour. Posteriorly, the upper part of the band fades gradually into a pale color, with (in pre- servative at least) silvery specks at the scale bases. These specks extend to the “scythe.” The greatest width of the stripe, where well defined, is about three-fourths that of the orbit. MIppoRSAL STRIPE ON HEAD. A third stripe, about one-third as wide as the orbit, dusky-centered and blackish-edged, extends from above the middle of the eve to the premaxillary groove. It is continued forward, on the top of the broad premaxillary expansion, as a solid blackish bar about one-fourth to one-sixth as wide as the orbit. SIDE OF HEAD. In advance of the “scythe” the sides of the head in life reflected rather strong blue glints, which were strongest on the opercles. The very edge of the orbit, posteriorly and ventrally, is h¢ht (bright yellow in life). Since the ocular bar is extended forward on the side of the snout there is no trace of a suborbital bar, and since the anterior part of the “seythe,” representing the second bar of many species, extends across the opercles, there are no markings about the pectoral base. In the field notes no evidence was recorded of definite light margins on the main dark bars, such as are evident on specimens of C. aya preserved in alcohol. ‘ ABOVE AND BEHIND THE “SCYTHE.” Above and behind the “scythe,” down to the level of the upper edge of the caudal peduncle, and extending over the thickened scaly base of the dorsal fin, the color in life was purplish- gray over a yellowish base. The purplish-gray remains in the alcohol-pre- served specimen. In this area there are faint dusky specks at the scale bases. Where it is exposed along the upper edge of and behind the “‘seythe,” VoL. XXIX] HUBBS &€ RECHNITZER: A NEW CHAETODON 293 the lateral line in life was iridescent silvery on the specialized seales, and in alcohol the line remains hght. In life the caudal peduncle was yellow-brown, becoming almost clear yellow ventrally. In aleohol the pedunele is lightened, especially on the lower surface. BELow THE “scyTHE.” Below the black mark the sides were brightened in life by silvery-violet reflections. In this area there were rather indistinct purplish-eray streaks along the middle of the scale rows. Anteriorly, this color became almost solid on the scales, except for a ereamy anterior border. The purplish-gray streaks remain in alcohol, and silvery flecks are generally interspersed between the streaks. VENTRAL SURFACE AND LOWER FINS. In life the ventral surface was lemon. This color extended onto the pelvie and anal fins and became intensified on the membranes behind the spines. Otherwise these fins were dusky olive-yvellow, shading outward to blackish-brown at the margin of the pelvie soft-rays (excluding the pale filament) and just within the margin of the soft part of the anal fin. In alcohol the anal spines are pale gray, but the soft anal becomes almost solid black on the sharp outer edge of the dark area. The clear margin becomes pinched out at the first anal soft-ray. Along the upper, posterior part of the fin the abruptly clear margin is about one-sixth as wide as the orbit. The pelvie in aleohol remains blackish toward the posterior margin, leaving the spine, the front of the first soft- ray, and the filament pale gray. The pectoral was essentially clear in life. In alcohol its upper margin is slightly darkened and the rays are set off by fine purplish edges. CaupAL FIN. In life this fin had some purplish-gray on the rays. In alcohol the fin is slightly dusky, with a blackish upper and lower border, and there are blackish edges along the rays near the middle third of their length. Dorsau FIN. In life the spines of this fin were purple-black and the median interspinal membranes were blotched with yellow and purplish-gray. The outer part of the fin from the seventh spine backward was sooty with some purple, and with some yellow showing through. The soft dorsal was purplsh-gray over pale vellow basally, then light yellowish in an irregular, subvertical band, which was followed by a blackish-purple band inside the almost clear, slightly purple-gray margin. In alcohol the blackish color remains. It is more or less solid on the exposed parts of the spines, except for a lighter basal streak from the third to the tenth spines. The first two membranes are almost solidly blackish. The third to sixth membranes are light, with blackish dappling. Except on the pale basal streak the follow- ine membranes, where exposed, are blackish, especially behind the spines. The blackish area on the outer-posterior part of the spinous dorsal is con- 294 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH Ser. tinued backward, and then downward on the soft dorsal, as a submarginal blackish band, which becomes narrower and intensified toward the end of the fin. On the anterior third of the soft dorsal the band is about as wide as the orbit and fades into the purplish base of the fin, but on the posterior two-thirds of the fin the blackish and purplish areas are separated by the remnants of the subvertical yellowish band. Throughout the fin the dark streak is abruptly separated from the margin by the clear light area, which, except toward the ends, is about one-third as wide as the orbit. Thus, the soft dorsal and the soft anal are very similarly colored. MEASUREMENTS PROPORTIONS OF PARTS COMPUTED AS THOUSANDTHS OF THE STANDARD LENGTH (138 mm.). Greatest body depth (from near pelvic insertion ver- tically to top of scaly sheath), 554; least depth of caudal pedunele, 104; length of peduncle from end of anal base to caudal flexure on midline, 86. Predorsal length, 425; prepelvie length, 483; caudal flexure to anal origin, 397; thence to pelvic insertion, 317; thence to isthmus, 223. Length of head to edge of opercular membrane, 354. Widths of head: at widest point, 130; at muzzle, across front of orbits, 81; across maxillaries, 70; opposite front of rostral groove, 45. Length of upper jaw, 103; midline length of upper hp, 46. From anterior nostril to rostral groove, 69. Length of orbit (be- tween free rims), 79. Least distance from free rim of orbit: to front of upper lip (length of snout), 147; to rostral fold (preorbital width), 97; to suborbital margin, behind mouth (suborbital width), 70; to other orbit (least fleshy interorbital width, anteriorly), 75; to farthest point on margin of opercular membrane (postorbital), 141; to preopercular angle, 119; to ventral contour of head, 109; to dorsal origin, 246. Length of dorsal spines (alone middle of side above edge of scaly sheath, disregarding sealy ex- tensions along front of alternate spines) : first, 96; second, 178; third and fourth, 280 each; fifth, 246; sixth, 225; seventh, 193; eighth, 168; ninth, 137; tenth, 120; eleventh, 100; twelfth, 85; thirteenth, 67. Length of anal spines, similarly measured from scaly sheath: first, 93; second, 198; third, 165. Other fin measurements: length of anal base, 310; length of middle caudal rays, 180; length of pectoral from upper axil, 266; width across pectoral base, 81; length of pelvic spine along front, 226; length of pelvie fin, including filament, 293. DERIVATION OF NAME The species name falcifer is the Latin word signifying seythe-bearer, in allusion to the diagnostic seythe-shaped color mark. We treat it as a noun and it is therefore not declinable. VoL. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 295 DISTINCTIVE CHARACTERS AND RELATIONSHIPS OF THE THREE OTHER EASTERN PACIFIC CHAETODONTINES An examination of the three other chaetodontines of the eastern Pacific, made in the University of California at Los Angeles by courtesy of Boyd W. Walker, has disclosed a considerable number of distinctive characters in the structure of the head and in the squamation of the head and caudal fin. Sharp differences between all four species are apparent in most of these characters. Some of these specific differences may well prove useful in the definition of species groups and subgenera, perhaps even of genera. Since we have no present opportunity to test the wider application of these features in the classification of the group, we offer only the following descriptive remarks, accompanied by inferences as to the relationships of the three species. Chaetodon humeralis Giinther. The series of specimens (W 51-20) studied in most detail was taken by Boyd W. Walker and party on January 25, 1951, on the shoreward side of Isla Venado, near Mazatlan, Sinaloa, México, but others were examined to verify the consistency of the characters. MoutH AND GAPE. As seen from above, the mouth is rather evenly arched and shows only a trace of lateral knobs. Lies. The upper lip is divided laterally by a prominent lengthwise eroove, behind which the skin is broken into rather fine but sharp folds and creases. The main lengthwise groove is continued backward and down- ward as a deep fissure, behind which the surface is sealy and before which it is smooth, except for an irregular subvertical groove seen only in this species. TEETH IN JAWs. In each jaw there are about 8 cross rows of fine teeth behind the somewhat enlarged outer row. In the lower jaw the rows are broadly curved. Bony MARGINS. The preopercle and the lower preorbital margin are finely denticulate. The bony orbital rim is weakly denticulate at the very edge only. There is no rough exposed bony area above the orbit. SQUAMATION ABOUT ORBIT AND MUZZLE. There is no scaleless area above the orbit. In fact, the narrow fleshy band along the upper part of the orbit is fully exposed and is closely covered with hard ctenoid scales. These scales expand the flattish interorbital area a little on each side. The scaleless fossa about the nostrils is very small. The snout is well scaled forward to the rostral groove except for a small area on each side about a slight knob 296 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. at the front of each preorbital. The lower edge of the preorbital is nearly concealed by scales. The posterior exposed lobe of the maxillary is covered with strongly ctenoid scales. SQUAMATION OF CAUDAL FIN. The caudal is covered by ctenoid seales over more than three-fourths of its length, leaving a naked margin some- what wider than that on the soft parts of the dorsal and anal. The basal band of body-type seales is broader than in the other species. The small though rough seales are in strips centered along the interradial membranes, but these strips contain several scale rows, and they are so wide that there intervenes between each pair only a narrow, more or less closed groove along the unbranched part of each ray out beyond the thickly sealed basal area of the fin. RELATIONSHIPS. This, the only previously described species from the eastern Pacific to be retained in Chaetodon, seems fully distinet from its nearest congenors in the Indo-Pacific and West Indian faunas. No attempt has been made to seek its nearest apparent relatives. Heniochus nigrirostris (ill). The specimen (W 52-258) examined in most detail was collected in Bahia San Lueas, Baja California, by Murray A. Newman and John E. Fitch, on December 1, 1952. Other examples were checked to verify the characters. MoutH AND Gare. The front of the gape is almost straight and trans- verse, with an almost lobular angle on either side of each jaw. Lips. In this species the lips are only weakly grooved in advance of, and above, about three moderate oblique fissures in the premaxillary region. There is no main longitudinal groove and no vertical groove on the maxil- lary, the exposed face of which is moderately rugose. TEETH IN JAWs. The inner teeth are in only about 6 transverse rows in the front part of each jaw. In the lower jaw the arrangement is less regular than in the upper jaw and the rows are less transverse, because of the close approximation of the mandibles. Bony MarGINS. The preopercle and the lower margin of the preorbital vary from nearly smooth to finely and regularly denticulate. The lower border of the orbit is finely denticulate with tiny uniserial teeth. On the upper orbital border the bony points are more numerous and are pluriserial. This rough and exposed bony margin continues into a triangular expansion above and inward from the middle of the orbit. On the nearly naked, slightly fleshy area on each side of the interorbital, inward from and in front VoL. XXIX |} HUBBS &€ RECHNITZER: A NEW CHAETODON 297 of the rugose area, there is evident a branching pattern of lateral-line canals. In the prefrontal region, the upper orbital margin is produced as a sheht flange, with somewhat strenethened denticulations. SQUAMATION ABOUT ORBIT AND MUZZLE. The fleshy upper orbital rim is sealeless in this species. Under the prefrontal! flange just mentioned this dermal band bears several finely papillate dermal ridges. The rugose bony area within and above the bony orbital rim is nearly all sealeless, as is a fleshy area in advance thereot. The scales on the top of the head become obsolete between the nostrils, with few extending slightly farther forward. There is a wide, almost wholly sealeless fossa about the nostrils, continuous with the naked front of the muzzle. There appear to be no seales on the maxillary. SQUAMATION OF CAUDAL FIN. The body-type seales form a rather nar- row basal band on the caudal, with extensions toward the upper and lower margins of the fin. Over most of the rest of the fin the seales (bearing mod- erate ctenil) extend much less than halfway to the rear margin; they are in single series overlying the subbasal part of the interradial membranes, leaving the intervening rays largely free of scales. GENERIC POSITION. As can be seen from the figures of nigrirostris pub- lished by Jordan and Evermann (1900: 3283, pl. 248, fig. 620), by Gilbert and Starks (1904: 148-149, pl. 24, fig. 47), and by Schultz (1951: 486, fig. 94), and as ean be determined from Hildebrand’s key and descriptions (Meek and Hildebrand, 1928: 766-770), nigrirostris differs sharply from other species referred to Chaetodon (including C. falcifer) im having the lateral line extended to the middle of the base of the caudal fin, instead of ending at the axil of the soft dorsal fin or at some point high on the fish anterior to and approximately in line with the axil. This distinction in posterior termination and position of the lateral line was used by Herre and Montalban (1927: 13-14), justifiably we think, as a primary character in the generic analysis of the family. Weber and de Beaufort (1936: 15-16) also utilized the character in their generic analysis of the Chaetodontinae (but, in this respect, inadvertently misplaced Heniochus in their key). The same or similar lateral-line characters have been employed in recent analyses of the genera of the subfamily Pomacanthinae (Fraser-Brunner, 1933; Smith, 1955). For these reasons we think that nigrirostris is misplaced in the genus Chaetodon. It does not, however, fit perfectly into any of the genera as defined by Herre and Montalban, or by such other revisers as Giinther (1860: 2), Bleeker (1878: 22-72, and preceding papers), Jordan and Ever- mann (1898: 1673; 1905: 362), Ahl (1927: 7), Weber and de Beaufort (1936: 14-16), and Woods (in Schultz, 1953: 566-575). 298 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. The species nigrirostris would fit into Heniochus quite well if the fourth dorsal spine were more produced or if the orbital or nuchal spines were developed (the apparent predorsal spine in Schultz’s 1951 figure 94 is a printer’s artifact, as can be seen by checking this figure with another, from the same drawing, published by Jordan and Evermann). But since Hen- ochus is characterized either by having the fourth dorsal spine more or less prolonged or by having hornlike processes on either the orbit or the nape, or both, nigrirostris may be interpreted as having been derived from the common ancestor of this assemblage, as thus heterogeneously diagnosed, for it, as the inferred ancestor presumably did, lacks the specialized processes and has the fourth dorsal spine only slightly produced (less so than in any species currently referred to Heniochus). Pending further studies, we sug- gest that nigrirostris be referred to Heniochus, and thus be named Heni- ochus nigrirostris (Gill). It might also be referred to Hemitaurichthys, though it differs from the species of that genus, as currently distinguished (as by Weber and de Beaufort, 1986: 15, 24-25), in having the fourth dorsal spine rather than the middle ones highest, the scales less reduced in size, and the body outline less elliptical. These are relatively minor char- acters. Until more trenchant differences are encountered, some doubt will pertain to the separation of Hemitaurichthys from Heniochus. We feel considerable reserve in the reference of nigrirostris to Heniochus, because the current classification of the chaetodontines is so unsatisfactory. Furthermore, Loren P. Woods, who has studied the family extensively, has suggested (in letter) that the contrasting lateral-line characters on which we have relied may tend to intergrade. But on checking his notes on this character, for 39 species that he placed in Chaetodon, it appears that he found the lateral line proper incomplete and ending far above the axis of the body in all, with the tubes ending at some point under the posterior half of the soft dorsal (or, under the middle part of the second dorsal in C. triangulum, which is referred to Gonochaetodon by Weber and de Beau- fort, and in C. auriga). In two species he found, in addition, some lateral- line structures in a separate series on the midside of the caudal peduncle (as is also true of some pomacanthines). In C. auriga Forskal he observed a few “pored seales” but no tubes on the pedunele, and in C. miliaris Quoy and Gaimard he located on the peduncle a double row of pitted scales. In the same two species we find only some neuromast pits (in part short-linear in ©. miliaris) in the axial scale row on the peduncle and on a few scales running upward and forward, but not nearly connecting with the lateral line proper. We have found similar neuromast pits in some specimens of C. aya, which may also have other pits on the first scale row above the lateral line and, scatteringly, elsewhere on the body. The type of C. falcifer has, on the right side only, two misplaced tubes and pores on the midside Vout. XXIX] HUBBS & RECHNITZER: A NEW CHAETODON 299 before the vertical from the end of the lateral line. We do not regard such irregularities as of any great taxonomic significance. Furthermore, Leonard P. Schultz, after examining for us a long range of species in the United States National Museum, concurs in our interpreta- tion of the generic significance of the lateral-line character and in the provisional reference of nigrirostris to Heniochus. Examination of many published figures seems to confirm the significance of the lateral-line dis- tinction. Certain apparent exceptions to the lateral-line character appearing on the plates in Bleeker’s “Atlas” (1878) become resolved when it is seen that the names and figures were inadvertently transposed on plate 375 for Coradion melanopus and Megaprotodon strigangulus and on plate 376 for Tetragonoptrus (Linophora) Rafflesii and Coradion chrysozonus. Forcipiger longirostris (Broussonet). The specimens studied were one (W 55-161) taken at Isla Clarion of the Revillagigedo group by Richard H. Rosenblatt and Raymond M. Gil- more on May 7, 1955, and two (W 53-351) from Isla Socorro, of the same group, taken by Bayard H. Brattstrom on November 18, 1953. MoutH AND GAPE. The very peculiar mouth of this species has been duly described. In top view it is very narrow and rounded. Lips. There are no prominent grooves or ridges on either the pre- maxillary or the maxillary region. TEETH AND JAWs. The inner teeth in the front part of the beak are arranged in about 20 to 25 cross rows in the upper jaw and in 25 or more rows in the lower jaw. Bony maArGINS. The preopercular edge varies from strictly entire to incipiently denticulate. The lower border of the preorbital is smooth ante- riorly, slightly to moderately denticulate posteriorly. The lower border of the suborbitals is either free (with smooth edge) or bound down. The upper and lower anterior corners of the preorbital are produced in bony lobes. When and where naked, the surface of the preorbital is more or less strongly sculptured, somewhat like the skull above each orbit. In that region, to a varying degree, the skull bones are exposed in a crescentie area, with definite bony ridges in a very complicated pattern, largely lengthwise anteriorly and dorsally, largely scroll-like posteriorly and ventrally. The actual bony rim is rough but hardly denticulate. SQUAMATION ABOUT ORBIT AND MUZZLE. The sealeless area above the orbital rim varies considerably, but near the rim is probably always naked. On the preorbital the scales may be eyeloid and confined to an anterodorsal patch, or may be ctenoid over almost the whole bone. There is a large 300 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH Ser. scaleless narial fossa and most of the top of the snout anterior to the nostrils is naked; but here also the completeness of the squamation varies. The top of the beak just before the rostral fold may be either naked, like the rest of the beak, or may bear a variable patch of smooth seales. The narrow fleshy band along the upper orbital rim is sealeless, and in some specimens is concealed beneath the bony rim. SQUAMATION OF CAUDAL FIN. The squamation of the tail fin is essen- tially as in H. nigrirostris, with a narrow basal band of ordinary overlap- ping seales and irregular uniserial rows extending out on the membranes, but ending short of the middle of the rays over most of the fin. RELATIONSHIPS. It seems highly probable that Forcipiger was derived from a Heniochus-like ancestor with a primitive lateral line. Its chief modi- fication seems to be the beaklike muzzle, which is presumably an adaptation for the procurement of food in small recesses, on the coral reefs that are spectacularly overgrazed. The interpretation of Prognathodes as convergent in the beaklike structure (see following section) lessens the apparent taxo- nomic value of the character and discounts any probable direct relationship between Forcipiger and such species of Chaetodon as C. falcifer and C. aya. TAXONOMIC NOTES ON ATLANTIC RELATIVES OF CHAETODON FALCIFER Following are the synonymies, records, and taxonomic notes for the three little-known Atlantic species that we regard as rather close relatives of Chaetodon falcifer. Chaetodon aya Jordan. (Plate II, figure 1; Plate III, figure 1.) Chaetodon aya JoRDAN, 1886: 225 (original description; distinguished by colora- tion; Snapper Banks, near Pensacola, Florida). EIGENMANN and HOoRNING, 1887: 5, 8, 18 (comparison and diagnosis, based on type; northern Gulf Coast of Florida; Pensacola). JoRDAN and EveRMANN, 1898: 1673, 1676-1677 (compari- sons; description; rather deep water; Gulf of Mexico). Aut, 1923: 122-123 (after Jordan and Evermann; placed in section Chaetodontops, misspelled Chae- dontops on p. 111). Breper, 1929: 215, 217 (comparisons; Gulf of Mexico; chiefly in rather deep water; “may represent the young of some other species’). eee PLATE II Figure 1. Chaetodon aya Jordan, from half-grown specimen (C.N.H.M. 45564) 65 mm. long, trawled off northern Florida at a depth of 25-30 fathoms. The dark margins on the soft dorsal and anal fins and the dark bar on the caudal are photo- graphic artifacts. Photo by Scripps Institution. Figure 2. Holotype of Chaetodon marcellae Poll, reproduced, by permission, from Poll (1950: fig. 1). VoL. XXIX | HUBBS & RECHNITZER: A NEW CHAETODON 301 Sean = 2 Sx FG. 2 302 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. Nicuots and Firru, 1939: 87-88 (off North Carolina; faunal relationships; description). HitpeBraNnpb, in LONGLEY and HILDEBRAND, 1941: 150-151 (Tortugas, Florida; depth distribution; color; counts and proportions; except smallest specimen, which is C. ocellatus). Increased collecting in deeper water has yielded a moderate amount of material of this species, which was long known only from the holotype. Following are data on the material in the United States National Museum and the Chicago Natural History Museum. No specimens were located in the Museum of Comparative Zoology, in the Academy of Natural Sciences of Philadelphia, or in the University of Michigan Museum of Zoology. The only other specimen known is the one reported by Nichols and Firth. U.S.N.M. 37747 (1 half-grown, 28.7 mm. in standard length, slightly over 114 inches to end of broken caudal): near Pensacola, Florida, Silas Stearns.—This is obviously the holotype. U.S.N.M. 37862 (1 subadult, 63 mm. long): “W. Coast of Florida, Str. Alba- tross, 1885.”—This specimen has not been mentioned in the literature. It is much too large to be the holotype. U.S.N.M. 116850 (7 specimens, half-grown to adult, 32-82 mm.): south of Tortugas, Florida——These are the specimens collected by William H. Longley and reported by Hildebrand (in Longley and Hildebrand, 1941: 150) as having been taken at two stations: the two largest, 70-82 mm. long, now tied together with tin tag number 12, from 39 fathoms; the five smallest, 32-69 mm. long, from 40 fathoms. These series correspond with Longley’s total-length measurements of 90-102 and 41-85 mm., respectively. The eighth specimen, 27 mm. in total length, 22 mm. to caudal, which Hildebrand reported as obviously representing a third collection of the species, for which no data could be found in Longley’s notes, proves to be a typical young example of Chaetodon ocellatus. It has been recataloged as U.S.N.M. 164519. U.S.N.M. 131893 (1 adult, 81 mm. long): Gulf of Mexico, at Albatross Sta. 2365, just north of the tip of Yucatan Peninsula, at 22° 18’ 00” N. Lat., 87° 04’ 00” W. Long., depth 24 fathoms, bottom of white rock and coral, January 30, 1885. U.S.N.M. 151980 (1 adult, 98.5 mm. long): Long Bay, South Carolina, trawled by the Albatross III on cruise 31-C, Sta. 3, tow 5 (No. 14593), from 33° 35.5’ N. Lat., 76° 53.5’ W. Long., to 33° 36’ N., 76° 51’ W., depth 40-45 fathoms, February 10, 1950.—This specimen is shown on plate III, figure 1. U.S.N.M. 152023 (1 adult, 82 mm. long): off Carolina Beach, North Carolina, trawled by the Albatross III on cruise 31-A, Sta. 2, tow 2 (No. 14506), from 34° 05’ N. Lat., 76° 21’ W. Long., to 34° 01’ N., 76° 18’ W., depth 25-75 fathoms, January 19, 1950. C.N.H.M. 45564 (2 subadults, 63-65 mm. long): Gulf of Mexico, off northern Florida, trawled by the Oregon at Sta. 727-728, 28° 44’ N. Lat., 85° 01’ W. Long., depth 25-30 fathoms, December 16, 1952.—One of these specimens is shown on plate II, figure 1. C.N.H.M. 46566 (1 adult, 77 mm. long): Gulf of Mexico, off southern Florida, trawled by the Oregon at Sta. 33, 25° 55’ N. Lat., 83° 53’ W. Long., depth 62 fathoms, June 24, 1950. VoL. XXIX | HUBBS &€ RECHNITZER: A NEW CHAETODON 303 C.N.H.M. 59898 (1 adult, 81.5 mm. long): Gulf of Mexico, off northern Florida, trawled by the Oregon at Sta. 916, 28° 23’ N. Lat., 84° 49’ W. Long., depth 37 fathoms, October 3, 1953. C.N.H.M. 64369 (1 adult, 75.5 mm. long): off North Carolina, trawled by the Combat at Sta. C-384, 35° 54’ N. Lat., 75° 25’ W. Long., depth 75 fathoms, June 17, US le Counts were made on all specimens examined, on both sides for bilateral structures. The variation found in scale counts is indicated in table 1. Following is the observed variation in ray counts, with number of specimens in parentheses: dorsal spines 13 (16), 14 (1); dorsal soft-rays, 17 (2), 18 (a eeto) (7); anal spmes, 3 (17): anal soft-ravs, 14 (2), 15 (14)0 16 (1): principal caudal rays, 16 (1, with fourth ray from bottom more widely forked than usual). 17 (16); pectoral rays, not distinguishing unbranched endunranched rays, 13—13) (il); 14-14 (10), 14—-2 (1), 14-15 (3), 215 Ch) aio to (CL))> pelvic rays, -5-5—I, 5°(17). Checking in sequence the description of C. falcifer, we noted the fol- lowing characters on the specimens of C. aya listed above. The outermost pelvic scft-ray is produced into a short filament, about as in C. falcifer. In some specimens a few neuromast pits were observed on the scales of the midlateral row on the caudal peduncle, and on a few seales near the base of the peduncle, in a row extending upward and slightly forward. Other seales, scattered, bear such pits. Some of the scales in the row impinging above on the lateral line bear pits in some individuals. The seales across the top of the premaxillaries in front of the rostral fold vary from none to a band covering about half the exposed width of the premaxillaries. The groove from the nostrils to the rostral fold is variously sealed over. Variable also is the degree of scalelessness and of bony tuberculation on the small triangu- lar area above and behind the bony orbital rim. The length of the gape varies from considerably less than half to a little more than half the length of the upper jaw. The bony ridge between the upper orbital border and the posterior nostril may be detached or scarcely developed, but may also extend nearly to the nostril. In most specimens the narrow fleshy upper orbital rim, more or less protruding from the bony rim, is devoid of scales, perhaps as the result of loss, because in some examples this margin bears spiny scales. In some individuals this fleshy rim is scarcely apparent. Especially in the fish preserved in aleohol, and therefore retaining guanin, the rear border or even both edges of the band above the eye and both borders of the poste- rior bar are abruptly margined by a light streak. In the squamation of the caudal fin C. aya may differ from C. falcifer, but the loss of seales on this fin is difficult to appraise. In a few specimens strongly etenoid scales cover most of the caudal fin, except along the posterior margin, but in general such seales are confined to the basal part of the fin and the scales that may 304 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH Ser. have been lost have not left any conspicuous pockets. Except as noted above or in table 1, and except for internal characters and detailed measure- ments that were not taken, the description of C. falcifer applies quite well to C. aya. Chaetodon marcellae Poll. (Plate II, figure 2.) Chaetodon (Oxychaetodon) marcellae Pott, 1950 (November): 2-7, fig. 1 (original description; comparisons; 25 miles southwest of Pointe de Banda, west Africa, at 3° 57.5’ S. Lat., 10° 36.5’ E. Long.). Chaetodon altipinnis CADENAT, “1950” (but on rear cover is indication of publica- tion in 1951, as follows: ‘‘dépot légal 1951, ler trimestre, no 236’’): 239, 307, 315, 318, fig. 174 (original diagnosis; Sénégal). This interesting species was described almost simultaneously by Poll and by Cadenat. Poll has advised us that his publication actually did ap- pear in 1950. It is obviously a very close relative of the western-Atlantic Chaetodon aya. In fact, the only really sharp difference that is apparent on comparing Poll’s description and figure with C. aya lies in the position of the posterior dark bar, which originates near the end instead of near the middle of the soft dorsal and is nearly vertical instead of rather strongly oblique (compare plate II, figure 1 and plate III, figure 1, with plate I, figure 2). This is obviously an adequate distinction. Poll referred marcellae to the Indo-Pacific subgenus Oxychaetodon, but the agreement is not close. He failed to appreciate the intimate relationship with C. aya, which until now has not been figured, and which Ahl referred to the subgenus Chaetodontops. Chaetodon marcellae, like C. aya and C. falcifer, is obviously a relatively deep-water species. The type was taken far offshore at a depth of 85 meters, on a bottom of sandy brown mud. This species may be included among the not inconsiderable number of essentially New World types that has become established in tropical west Africa (p. 308). Prognathodes aculeatus (Poey). (Plate III, figure 2.) Chelmon aculeatus Pory, 1860 (July): 202-203 (original description; rare; Cuba). Prognathodes aculeatus Porky, 1868: 354 (reference; characters). EIGENMANN and HorRNING, 1887: 2-3 (diagnosis; relations; after Giinther and Poey). JORDAN and EveERMANN, 1898: 1671 (description; synonymy; after Giinther and Poey). AHL, 1923: 11 (synonomy; relations; description; West Indies; Havana, Cuba). Prognathodus aculeatus BLEEKER, 1876a: 3038, 1876b: 315, and 1877: 32, 34 (charac- ters; genus; Synonymy). Chelmo pelta GUNTHER, 1860 (September): 38 (original description; locality un- certain). VoL. XXIX } HUBBS & RECHNITZER: A NEW CHAETODON 305 As the synonymy indicates, this species was originally associated with Chelmon. Ever since the genus Prognathodes was based on this species by Hl (1862: 238), it has been regarded as a close relative of the Indo-Pacifie genera Chelmon, Chelmonops, and Forcipiger, obviously on the basis of the beakhke modification of the muzzle in all four genera. But those three genera have the lateral line of the Heniochus type (continuous to the caudal base), whereas in Prognathodes aculeatus the lateral line ends, as in Chae- todon aya, under the anterior part of the soft-dorsal fin, only about one scale row from the structural base of the rays. Since we regard the lateral- line character as of prime taxonomic significance, we interpret Chelmon, Chelmonops, and Forcipiger as modifications of a primitive member of the Heniochus series, and Prognathodes as only a parallel modification of Chaetodon. In fact, Prognathodes aculeatus appears to have arisen from Chaetodon aya or from a very similar, related species. This view was anticipated by Hildebrand in his estimate of C. aya as being rather intermediate between more ordinary species of Chaetodon and Prognathodes. The many points of resemblance between Prognathodes aculeatus and Chaetodon aya can hardly be fortuitous. In its beaklike modification P. aculeatus merely exag- gerates the characters of C. aya. In squamation, body form, extremely strong and greatly excised fin spines, and other respects, the two species are remarkably alike (see plate II). Even the fin formula is approximately the same (the type of P. aculeatus has XIII, 19 dorsal, III, 15 anal, 17 eaudal, 14—14 pectoral, and I, 5 — I, 5 pelvic rays. Each species has finely denticulate bones around the orbit and has the muzzle well scaled forward to the rostral edge, with a more or less sealeless groove from nostril to the rostral fold, and the chin well scaled almost to its front. Each has the pre- opercular and scapular edges finely denticulate. In each the lateral line ends under the anterior part of the soft dorsal, as is noted above. The dark band through the eye is very similar. Although the two forms are so much alike, there can be no doubt that P. aculeatus is at least specifically distinet from C. aya. It differs from C. aya not only in the much more modified beak, as described below, but also in other respects. For instance, the posterior dark bar is apparently lacking, and the caudal peduncle is slenderer (ieast depth is less instead of more than one-fourth the leneth of the head). In view of the definite approach to Prognathodes of some species of Chaetodon, notably C. falcifer, C. aya, and C. marcellae, some ichthyologists may not wish to retain Prognathodes as a distinet genus. The only known differences lie in the beak structure. The jaws are slenderer and more pro- duced. The upper lp instead of covering much of the premaxillaries is confined to the extreme edge, leaving nearly all of the exposed premaxillary 306 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. Vou. XXIX] HUBBS &€ RECHNITZER: A NEW CHAETODON 307 surface smooth, bony, and sealeless, rather than rough, fleshy, and more or less scaled. Pending a more general study of the whole group, we suggest that Prognathodes be retained as a monotypic genus. The only specimen of P. aculeatus that we have located is U.S.N.M. 4716. It is labelled as from ‘Cuba, Prof. F. Poey,” and it bears Felipe Poey’s original number 56. Although it seems slightly smaller than the specimen described by Poey (61 mm. in standard length, 73 mm. to the somewhat broken tip of the caudal, rather than 80 mm.) it is regarded as the holotype of Chelmon aculeatus Poey, and has been so designated in the National Museum records (Poey’s measurement may have been rounded out to 80 mm. and the specimen may have shrunk slightly since he measured it; restored, the total length would probably slightly exceed 75 mm.). The specimen was entered as Prognathodes pelta and is no doubt the one Gill examined when he described the genus, designating P. pelta as the type species. It is shown as plate ITI, figure 2. SYSTEMATIC AND ZOOGEOGRAPHICAL CONCLUSIONS Chaetodon falcifer is described as a strikingly distinct new species of butterflyfish from rather deep water at Guadalupe Island, an oceanic island in the Pacific Ocean well off the coast of Baja California. It is regarded as probably an endemie species, in an island fauna marked by a high inci- dence of endemism. It is, however, a deep-water form, which may have escaped attention elsewhere. Whether endemie or not on Guadalupe Island, C. falcifer ranks as one of the tropical elements in the mixed tropical and temperate fauna of this island. Currently, it is probably the northernmost representative of the Chaetodontidae in the eastern Pacific. There are apparently reliable indi- cations, however, that Chaetodon humeralis ranged north to San Diego during the warm period approximately one hundred years ago. The closest relatives of C. falcifer appear, on the basis of similar mor- phology and of the distinctively deep-water habitat, to be two Atlantic spe- cies, C. aya of the western North Atlantie and C. marcellae of tropical west Africa. The West Indian Prognathodes aculeatus appears to be a derivative of C. aya (or of a closely related species), differing in the more extreme modification of the beak. These circumstances illustrate two zoogeographieal PLATE III Figure 1. Chaetodon aya Jordan, from adult (U.S.N.M. 151980) 98.5 mm. long, trawled off Long Bay, South Carolina, at a depth of 40-45 fathoms. Photo by United States National Museum. Figure 2. Holotype of Chelmon aculeatus Poey = Prognathodes aculeatus (Poey), a specimen (U.S.N.M. 4716) 61 mm. in standard length. Photo by United States National Museum. 308 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. tendencies that are being increasingly indicated by critical taxonomic studies; namely, (1) the tendency for New World fishes to possess a marked inde- pendence and integrity, indicative of lines of evolution essentially distinct from those of related Old World types, and (2) the tendency for New World types to be sparingly but definitely represented in the fauna of tropical west Africa. In the classification of the Chaetodontidae far too much emphasis has been placed on the beaklike modification, such as is exhibited to a moderate degree by the species C. falcifer, C. aya, and C. marcellae, and, much more strikingly, by Prognathodes aculeatus. Thus, P. aculeatus has been con- sistently associated with the recognized genera Chelmon, Chelmonops, and Forcipiger because of the similar beakline feature, but those genera have a more primitive and more complete lateral line and appear to have stemmed from the Heniochus group, whereas Prognathodes is rather clearly a deriva- tive of Chaetodon. Genera in several other families of coral-reef fishes, for example, Lo among the Siganidae and Gomphosus among the Labridae, have developed a similar beak, which seems to be a modification for plucking food out of generally inaccessible crevices, on the notably overgrazed coral reefs. This is one of many examples among fishes of a nutritional character that has been overemphasized in taxonomy, but which is much subject to parallel and convergent evolution. It seems rather clear that the beak was independently evolved in the chaetodontines with an incomplete lateral line and in those with a complete line. The species Chaetodon falcifer, C. aya, C. marcellae, and Prognathodes aculeatus probably constitute one such phyletie line, but other species of Chaetodon with an incipient beak, such as those referred to subgenus Oxy- chaetodon, are likely of separate origin. Of the beaked chaetodontines with a complete lateral line, we see no assurance that Chelmon, Chelmonops, and Forcipiger represent a single natural group. Certainly Bleeker, Ahl, Jor- dan, and other students have erred in classing together all chaetodontids with a beak. The lateral-line character appears to be of greater significance as an index of phylogeny in the group. On the basis of its simpler, more complete and more primitive lateral line, the eastern Pacific species Chaetodon nigri- rostris is referred to the Heniochus series, tentatively to the genus Heni- ochus. Heniochus ngrirostris and Forcipiger longirostris are the only New World members of this series, and in the western hemisphere they are con- fined to the Pacific. They presumably represent, respectively, relatively old and new migrants, from the vast Indo-Pacific fauna, that have crossed the east-Pacific barrier. The Chaetodontinae as a whole are chiefly Indo- Pacific. On these interpretations the high specialization of most species of Heni- VoL. XXIX] HUBBS & RECHNITZER: A NEW CHAETODON 309 ochus and of Chelmon, Chelmonops, and Forcipiger is coupled with a basically primitive feature in lateral-line structure. This view is consistent with the increasingly evident generalization that specialized and primitive characters tend to be combined. Various anatomical details will apparently prove of value in the much- needed generic revision of the family. Some characters of this sort are pointed out for the four species of chaetodontines now known from the eastern Pacific. The lateral-line distinctions are repeated in the Pomacanthinae, and it is conceivable that they may be of greater phylogcnetic significance than the enlarged head spines recognized as diagnostic of that subfamily. The pri- mary divergence in the family may have involved the lateral-line character, and each division may have developed one or more enlarged head spines. If this view should prevail, the Pomacanthinae would appear to be a poly- phyletic group, presumably not worthy of subfamily recognition. We see no justification for the elevation of the group to family rank (Pomacan- thidae), as proposed by Smith (1955), even though we assume it to be a natural assemblage. Obviously there has been independent and convergent evolution either in the head-spine or lateral-line characters. On available evidence a decision as to which character is of more primary significance seems arbitrary. Such a situation is often encountered when taxonomic evidence in ichthyology is viewed afresh and critically. LITERATURE CITED AHL, ERNST 1923. Zur Kenntnis der Knochenfischfamilie Chaetodontidae, insbesondere der Unterfamilie Chaetodontinae. Archiv fiir Naturgeschichte, Abt. A, 89 (5) :1-205, pls. 1-3. BLEEKER, PIETER 1876a. Systema percarum revisum. Pars II. Archives Néerlandaises des Sci- ences exactes et naturelles, sér. B, Sciences naturelles, 11: 289-340. 1876b. Notice sur les genres et sur les espéces des Chétodontoides de la sous- famille des Taurichthyiformes. Verslagen en Mededeelingen der Kon- inklijke Akademie van Wetenschappen, Amsterdam, Afdeeling Na- tuurkunde, ser. 2, 10(3) :308-320. 1877. Révision des espéces insulindiennes de la famille des Chétodontoides. Verhandelingen der Koninklijke Akademie van Wetenschappen, Am- sterdam, Afdeeling Natuurkunde, 17:1-174. 1878. Atlas ichthyologique des Indes Orientales Néerlandaises. Vol. 9:1—80, pls. 355-420. Fréderic Muller, Amsterdam. 310 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH Ser. BiLocH, MARC ELIESER 1796. Ichthyologie ou histoire naturelle des Poissons. Part 4: 969-1291; part 6:pls. 109-216. Berlin. BREDER, CHARLES M., JR. 1929. Field book of marine fishes of the Atlantic coast from Labrador to Texas. i-xxxvii, 1-332, many figs. G. P. Putnam’s Sons, New York and London. Brices, JOHN C. 1955. A monograph of the clingfishes (order Xenopterygii). Stanford Ich- thyological Bulletin, 6:i—iv, 1-224, figs. 1-114, maps 1-15. CADENAT, J. “1950” (1951). Poissons de mer du Sénégal. Initiations africaines, III. 1-345, figs. 1-241. Institut francais d’Afrique Noir, Dakar. CuvigeR and VALENCIENNES 1831. Histoire naturelle des Poissons. Vol. 7:i-xxix + 5 + 1-531, pls. 170-208. Paris. EIGENMANN, CArt H., and JENNIE E. HORNING 1887. A review of the Chaetodontidae of North America. Annals of the New York Academy of Sciences, 4(7—-8) :1-18. EKMAN, SVEN 1953. Zoogeography of the sea. i-xiv, 1-417, figs. 1-121. Sidgwick and Jackson, Limited, London. FOWLER, HENRY W. 1928. The fishes of Oceania. Memoirs of the Bernice P. Bishop Museum, 10: i-iii, 1-540, figs. 1-82, pls. 1-49. Fow ter, HENRY W., and Barton A. BEAN 1929. The fishes of the series Capriformes, Ephippiformes, and Squamipennes, collected by the United States Bureau of Fisheries Steamer ‘“‘Albatross,” chiefly in Philippine seas and adjacent waters. United States National Museum, Bulletin 100, Vol. 8:i—xi, 1-352, figs. 1-25. FRASER-BRUNNER, A. 1933. A revision of the chaetodont fishes of the subfamily Pomacanthinae. Proceedings of the Zoological Society of London, 1933:543-599, figs. 1-29, pl. 1. 1945. On the systematic position of a fish, Microcanthus strigatus (C. & V.). Annals and Magazine of Natural History, ser. 11, 12(91):462—468, figs. 1-3. GILBERT, CHARLES H., and Epwin C. STARKS 1904. The fishes of Panama Bay. Memoirs of the California Academy of Sciences, 4:1-304, pls. 1-33. VoL. XXIX] HUBBS & RECHNITZER: A NEW CHAETODON 311 GILL, THEODORE 1862. Remarks on the relations of the genera and other groups of Cuban fishes. Proceedings of the Academy of Natural Sciences of Philadelphia, [{ vol. 14] 1862 (5) : 235-242. GIRARD, CHARLES 1858. Fishes. General report upon the zoology of the several Pacific railroad routes. Part 4:i-xiv, 1-400, 21 pls. Jn Reports of explorations and surveys ... for a railroad from the Mississippi River to the Pacific Ocean, vol. 10. U. S. Senate Ex. Doc. no. 78, 33d Congress, 2d Session, Washington. GUNTHER, ALBERT 1860. Catalogue of the acanthopterygian fishes in the collection of the British Museum. Squamipinnes, Cirrhitidae, Triglidae, Trachinidae, Sciaeni- dae, Polynemidae, Sphyraenidae, Trichiuridae, Scombridae, Carangidae, Xiphiidae. Vol. 2:i-xxi, 1-548, 1 fig. London. 1868. An account of the fishes of the states of Central America, based on col- lections made by Capt. J. M. Dow, F. Godman, Esq., and O. Salvin, Esq. Transactions of the Zoological Society of London, 6(7):377—494, pls. 63-87. HERRE, ALBERT W., and HERACLIO R. MONTALBAN 1927. The Philippine butterfly fishes and their allies. The Philippine Journal of Science, 34(1):1-113, pls. 1-24. Huss, Cart L. 1946. An arm protractor for the precise measurement of angles in systematic ichthyology. Copeia, 1946 (2):79-80, fig. 1. 1948. Changes in the fish fauna of western North America correlated with changes in ocean temperatures. Journal of Marine Research, 7(3): 459-482, figs. 1-6. Husss, Cari L., and Kari F. LAGLER 1952. Fishes of the Great Lakes region. Cranbrook Institute of Science, Bul- letin 26:i—xi, 1-186, figs. 1-251 4 38, 26 col. pls., 1 map. JORDAN, DAvip S. 1886. Notes on some fishes collected at Pensacola by Mr. Silas Stearns, with descriptions of one new species (Chaetodon aya). Proceedings of the United States National Museum, 9:225-229. JORDAN, DAvip STARR, and BARTON WARREN EVERMANN 1898. The fishes of North and Middle America: a descriptive catalogue of the species of fish-like vertebrates fecund in the waters of North America, north of the Isthmus of Panama. United States National Museum, Bulletin 47(2) :i-xxx, 1241-2183, 1 fig. 1900. The fishes of North and Middle America. ... United States National Mu- seum, Bulletin 47 (4) :i-ci, 3137-33138, pls. 1-392. 312 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. JORDAN, DAvip STARR, and BARTON WARREN EvERMANN—Cont. 1905. The shore fishes of the Hawaiian Islands, with a general account of the fish fauna. Bulletin of the United States Fish Commission, for 1903, 23 (1) :i-xxviii, 1-574, figs. 1-229, pls. 1-65, col. pls. 1-73. JORDAN, DAviIpD STARR, and HENRY W. FowLer 1902. A review of the Chaetodontidae and related families of fishes found in the waters of Japan. Proceedings of the United States National Mu- seum, 25:513-563, figs. 1-6. JORDAN, DAVID STARR, and ALVIN SEALE 1906. The fishes of Samoa. Description of the species found in the archipelago, with a provisional check-list of the fishes of Oceania. Bulletin of the [United States] Bureau of Fisheries, [for] 1905, 25:173-455, figs. 1-111, pls. 33-37, col. pls. 38-53. KENDALL, WILLIAM C., and EpmuNpD L. GOLDSBOROUGH 1911. Reports on the scientific results of the expedition to the tropical Pacific, in charge of Alexander Agassiz, by the U. S. Fish Commission Steamer “Albatross,” from August, 1899, to March, 1900, Commander Jefferson S. Moser, U.S.N., commanding. XIII. The shore fishes. Memoirs of the Mu- seum of Comparative Zodlogy at Harvard College, 26: 241-344, pls. 1-7. LONGLEY, WILLIAM H., and SAMUEL F. HILDEBRAND 1941. Systematic catalogue of the fishes of Tortugas, Florida, with observations on color, habits, and local distribution. Carnegie Institution of Wash- ington Publication 535, Papers from the Tortugas Laboratory, 34: i—xiii, 1-331, pls. 1-34. MEEK, SETH E., and SAMUEL F.. HILDEBRAND 1928. The marine fishes of Panama. Field Museum of Natural History, Publi- cation 249, Zool. ser., 15(3) :xxv—xxxi, 709-1045, pls. 72-102. Munro, IAn_S. R. 1955. The marine and fresh water fishes of Ceylon, i-xvi, 1-351, figs. 1-19, pls. 1-56. Department of External Affairs, Canberra. MYERS, GEORGE S. 1941. The fish fauna of the Pacific Ocean, with especial reference to zodgeo- graphical regions and distribution as they affect the international aspects of the fisheries. Proceedings of the Sixth Pacific Science Con- gress of the Pacific Science Association Held at the University of Cali- fornia, Berkeley, Stanford University, and San Francisco, July 24th to August 12th, 1939, 3, 1940’: 201-210. NIcHo is, J. T., and F. E. Firru 1939. Rare fishes off the Atlantic coast, including a new grammicolepid. Pro- ceedings of the Biological Society of Washington, 52:85-88, fig. 1. POEY, FELIPE 1860. Memorias sobre la historia natural de la isla de Cuba, acompafiadas de sumarios latinos y extractos en francés. Vol. 2(3):97-336, pls. 10-12, 14, Habana. VoL. XXIX | HUBBS &€ RECHNITZER: A NEW CHAETODON 313 Pory, FELIPpE—Cont. 1868. Synopsis piscium cubensium. Repertorio fisico-natural de la isla de Cuba, 2:279-484 (reprinted, in same year, with same pagination and with added title, Catalogo razonado de los peces de Jla isla de Cuba). POLL, MAx 1950. Description de deux Poissons percomorphes nouveaux des eaux cotieres de l’Atlantique Sud (1948-1949). Institut royal des Sciences naturelles de Belgique. Bulletin 26(49) :1—-14, figs. 1-2. SCHULTZ, LEONARD P. 1951. Chaetodon tinkeri, a new species of butterflyfish (Chaetodontidae) from the Hawaiian Islands. Proceedings of the United States National Mu- seum, 101:485—488, fig. 94, pl. 15. Smi1TH, J. L. B. 1955. The fishes of the family Pomacanthidae in the western Indian Ocean. Annals and Magazine of Natural History, ser. 12, 8(89):377-384, pls. 4-5. STEARNS, SILAS 1887. The fishing-grounds of the Gulf of Mexico belonging to the United States. 55-58, charts 16-17. In Fisheries and Fishery Industries of the United States, by George Brown Goode and associates. Sect. 3. United States Commission of Fish and Fisheries, Washington. Weber, MAx, and L. F. pre BEAUFORT 1936. The fishes of the Indo-Australian Archipelago. Perciformes (continued). Families: Chaetodontidae, Toxotidae, Monodactylidae, Pempheridae, Kyphosidae, Lutjanidae, Lobotidae, Sparidae, Nandidae, Sciaenidae, Malacanthidae, Cepolidae. Vol. 7:i-xvi, 1-607, figs. 1-106. E. J. Brill, Leiden. Woops, LoREN P. 1953. Subfamily Chaetodontinae. 566—596, pls. 49-58. In Fishes of the Marshall and Marianas islands: Asymmetrontidae through Siganidae, by Leonard P. Schultz and collaborators, United States National Museum, Bulletin 202, vol. 1. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Fourth Series Vol. XXIX, No. 9, pp. 315-360, 4 figs. January 5, 1959 BIBLIOGRAPHY, CARTOGRAPHY, DISCOVERY, AND EXPLORATION OF THE ISLAS REVILLAGIGEDO' BY ADRIAN F. RICHARDS” Scripps Institution of Oceanography, La Jolla, California and BAYARD H. BRATTSTROM Department of Biology, Adelphi College, Garden City, New York TABLE OF CONTENTS Page Mite ALONG Ste osha oer cc's 56 5 0i-6s, Ges een RSE ne ee ear ae 316 PAO orAp Ayana MOMeENClALUTE. « «4 4ac: ch » yg Glia a. .. ee 7 av" $08 et eee wie shed 2. Aah Pat) Jo ae : i= ti ty <@\ Gee ee rs TH me, A e wr - - a . _ 1p 4 ae peace. > . 20a \d Ate Be :4 ‘. >_> = ~_ i — a 7wen = PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES FOURTH SERIES Vol. XXIX, No. 11, pp. 423-444, 10 figs. February 27, 1959 TWO NEW SPECIES OF WEST NORTH AMERICAN MARINE GASTROPODS BY RUDOLF STOHLER Department of Zoology University of California, Berkeley Early in April, 1957, two shells were submitted to me for identification, one by Mrs. C. D. Jennings of Cardiff by the Sea, and the other by Mrs. A. Wolfson of San Diego, California. Information supplied at the time indicated that the specimens were taken alive in about 70 feet of water some time in December, 1956, at the Coronado Islands by Mr. Cecil D. Jennings. It was further stated that they had been heavily encrusted and that each of the two ladies had spent several hours cleaning her specimen. The two shells undoubtedly belonged to Astraea, and more particularly to the subgenus Uvanilla (Keen, 1956), but differed from any of the west- coast species known. These facts, together with the information supplied by Mr. Jennings to the effect that the two shells were taken less than two feet from each other, suggested several problems. In an attempt to solve these, a determined effort was made to find the original locality and to obtain additional specimens, including some younger stages. This was done during the month of July, 1957, in cooperation with Dr. E. W. Fager, associate professor of marine biology, Scripps Institution of Oceanography, La Jolla, and his close associate, Mr. Ray Ghelardi. The search comprised a large number of dives made at a great many stations along the shores of [423] 424 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47H SER. the Coronado Islands which lie off the coast of Lower California. The rich material obtained included at least one new species of marine gastropod and which is deseribed herein, but the sought-for Astraea was not found. Through a process of elimination, the log of dives made by Mr. Jennings revealed that one additional locality remained to be explored. This was situated about eight miles southeast from the south tip of South Coronado Island, a place called the “Rockpile” by the divers. This is a submarine plateau, reaching up to about 70 feet below sea level at its higher central area, sloping more or less gradually down to about 110 or 120 feet below sea level at its periphery, and then dropping off rather steeply and sud- denly. The plateau is estimated to be about a half mile in diameter. On July 26, 1957, six divers on four successive dives explored this area; Dr. Fager located the third specimen known (our paratype IT), crawling among the rocks at a depth of 110 feet. This animal was heavily encrusted, as has been described for the original two (the holotype and paratype I). An attempt was made to bring this new specimen back alive to compare it with living examples of Astraea undosa (Wood) and to study it in general; it was placed, therefore, in a live-bait tank aboard the ship, but for unknown reasons did not survive the trip back to base. On November 19, 1957, Dr. Fager, with four diving companions, again explored the “Rockpile” but limited the search to the higher area of the plateau; this time one young and six adult specimens were obtained. Dr. Fager brought the adult animals by air to Berkeley on November 22. Im- mediately, upon arrival, they were placed in well aerated aquaria; five of these animals remained alive for from five weeks to two months (paratypes III to VIT, below) while the sixth specimen is still living at the time this report is written (January 29, 1958). The young individual was preserved and is not included in the type material of the present study. The finding of this many individuals, all agreeing very closely, except for small individual variations, conclusively demonstrated that the first two shells are not just chance variants of Astraea undosa, but constitute some- thing different. In comparing the shells and their opercula with the type figure of Astraea petrothauma Berry, described in 1940, the possibility arose that these shells were living representatives of a species previously known only from its fossil record. This was especially true when the operculum was compared with figure 3 on plate 25 (Berry, I. c.). While a comparison of our shells with that of the paratype of A. petrothauma in the collection of the Department of Paleontology at Stanford University seemed to show a sufficient number of points of difference between the two, the final decision to consider the animals from the “Rockpile”’ as members of a heretofore unknown species was arrived at only upon finding still another shell. In Vou. XXIX | STOHLER: TWO NEW SPECIES OF GASTROPODS 425 August, 1957, Dr. Wheeler North of Seripps Institution of Oceanography, with a group of assistants, conducted a survey of the kelp beds off the coast of Lower California; at various stations he gathered gastropods for the collection of the Department of Zoology in Berkeley. This material finally reached me in early January, 1958. Among the various lots was one from Sacramento Reef, near Geronimo Island, Lower California, containing a dif- ferent form of Astraea, which agrees so completely with Berry’s description, figure, and the paratype at Stanford University, that there is little doubt about its being properly labeled as A. petrothawma Berry. This specimen demonstrated clearly enough the differences existing between Berry’s and the new species. After the manuscript for this paper was submitted, but before it could appear in print, one further living member of Astraea petrothauma Berry was found. A SCUBA training class, under the supervision of Mr. Conrad Limbaugh, on June 27, 1958, was taken to South Coronado Island. Mr. Limbaugh, being aware of my special interest in Astraea, collected 36 young animals, all of which proved to belong to Astraea undosa (Wood) except for one specimen. The series of A. wndosa obtained will be of great value in a planned further study of the entire group of species of this genus occurring on the west coast of North America. But of greater immediate importance was that other specimen. Unfortunately the animal did not survive the trip back to the laboratory, so that still no observations of a living representative of this very interesting species could be made. It seems, however, to add to the conviction that the specimens from the “Rock- pile” are sufficiently different to warrant their being assigned to a new species. The reader is referred to table 3 below for the measurements of this second specimen of A. petrothawma. Astraea (Uvanilla) rupicollina Stohler, new species. SHELL: Large, turbinate (fig. 1), not umbilicated, white, covered with a brown (Maerz and Paul, plate 14, G6)", lamellose periostracum; whorls six, rounded, with two distinct carinae of equal prominence on the body whorl, but the upper carina stronger than the peripheral carina on the earlier whorls (fig. 2c); 23 nodules each on both carinae on the body whorl, the 23 oblique, large ribs above the carinae showing a tendency toward dis- solving into separate nodules; base rounded, with three distinct concentric corrugations of which the outer is close to the periphery and nearly three 1 Colors were identified, whenever possible, with the aid of the dictionary of color by Maerz and Paul, 1930, and are quoted in a manner which is thought self-explanatory; because of the impossibility of matching colors of small parts of the living snail body with the plates, such colors are given subjective names and are not followed by references to the dictionary. 426 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47H SER. f Gs pocorn NS Sty ee An (LAN Ws, \\ \ As ‘ NYY js» WOON we ‘ rig 5) p tes SSS BARS SA = QRS \ LY SA SQW SS \\ SSQRSARMANS SS SQr NESS \ AW CRA \ AS ANY \\ \\ i \w) WS EARN) cc SEW WX TOS SYN WHEN \ SY : SS &\\: \\\ : \ US I >» \ LTRS RENESN ff MASS LBW S MWK ANA N\ \ N ‘ N WK QYAAY LAA f! EQUI QYAY \ . XAssy \ AAS \ \S ; APA 6 vq Wor Sas cy WY \ \\ {x Wn \\ QQ A\\\S WV RA er We i\ WAY \ WY OO SANT: as \\ \ ‘ AWN QA SN ee - : = \\ws WAX SS Fig. 1. Astraea rupicollina, new species. Front view of holotype showing opercu- lum withdrawn and in its proper position. VoL. XXIX] STOHLER: TWO NEW SPECIES OF GASTROPODS 427 | \ \} ue . ithe Fig. 2. Outlines of the three species of Astraea compared with each other to show the relative strength of the carinae and the fact that the nodules may be hollow or filled in. a=A. undosa (Wood) with hollow nodes; b=A. petrothauma Berry with hollow nodes and equal carinae; c= A. rupicollina, new species, with solid nodes and the secondary carina stronger than the primary one. times as far distant from the middle corrugation as the latter is from the inner corrugation (fig. 3); there is a faintly perceptible trace of a fourth corrugation about equidistant between the peripheral and the middle corru- gation; the peripheral corrugation shows a tendency toward nodulose sub- division; aperture round, outer lip thin, wavy, inner lip heavy, nacreous; inside of aperture pearly; inside of nodules on body whorl not hollow (cf. figs. 2a to 2c); columella semicircular, accompanied by a semilunar, fairly deep depression, the outer boundary of which in the umbilical area flattens out into the general callus area, but is produced into a low ridge toward the base of the crescent; still further basally is an additional depression, rela- tively small and bordered by a heavy, blunt, rounded ridge; a large portion of the base is covered with a thin, nacreous, smooth callus which seems to continue into an advanced incremental area where the periostracum of the shell is dissolved and the fine ribbing of the shell, corresponding to the lamellae of the periostracum, is not yet covered with the new callus material, giving this area a worn appearance; periostracum apparently worn away 428 CALIFORNIA ACADEMY OF SCIENCES [ Proc, 4TH SER. ] WSS ——— \ \y \\\ Y ) y SSE: Sn \\ ) Wh » i) NG | Ni ! /; ‘ZG Fig. 3. Astraea rupicollina, Fig. 4. Astraea rupicollina, new species. Operculum of the holotype. a: outside view; b: side view; c: muscle attachment or periostracum side. VoL. XXIX | STOHLER: TWO NEW SPECIES OF GASTROPODS 429 on the underside of the first seven to eight nodes of the lower carina and on the peripheral corrugation immediately adjacent to the eallus. OpERCULUM: Heavy, broad-ovate with three spirally curved, prominent ribs (fig. 4), bearing fine pustulate rugosities on the outer surface, rapidly diminishing in number from the nuclear to the distal end of the ribs and lacking completely on the distal five-sixths of the high, median, overhanging rib, as well as on the distal half of the next longest (outer) rib, which is much lower than the median rib; rugosities on the inner rib almost obsolete on the entire length; outer surface of the operculum completely smooth without trace of a lip; nuclear area recurved, tinged with brown; otherwise the outer surface of the operculum is white, shiny; opercular periostracum limited to the flat muscle attachment (inner) side, horny, heavy, sculptured with low, flat, spirally curved ribs and transversely curved, somewhat stronger, more or less concentric incremental ridges; the spiral ribs are noticeable even in the nuclear portion of the operculum; color of the nuclear area reddish brown (Maerz & Paul, plate 6, D 12), the distal portion yellow (M. & P., pl. 10, J 1) with the ribs and incremental ridges a slate blue (M. & P., pl. 48, C 11) (see fig. 5a). ANIMAL: Approximately the lower half of the sides of the foot is of a pre- dominantly port-wine color (M. & P., pl. 56, J 12), while the upper half is close to red brown (M. & P., pl. 7, H 8); epipodial tentacles four, pure white, short, triangular, gradually diminishing in size from the anterior to the Fig. 5. (a) Operculum of the holotype of Astraea rupicollina, new species, show- ing the periostracum. (b) Operculum of Astraea undosa (Wood). 430 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. Fig. 6. Astraea rupicollina, new species, living animal. Note the four epipodial tentacles. (This is a composite picture, carefully superimposing the cleaned holotype shell upon the heavily encrusted shell of the still living animal, from which the shape and pattern of the soft parts were painted. posterior end of the row; mantle fold pure white, rimmed with a narrow, jet black margin, which in turn is bordered in the portion lining the aperture by a delicate yellow tinge; the mantle fold also surrounds and partly covers the operculum when the animal is active, outlining the operculum with the jet black margin (see figs. 6 and 7); eyes on prominent, short stalks; tentacles slender, pointed, short, about twice as long as the eye stalks. Rapuua: Broad, flat, long, with four rows of plates; lateral plates each ending in ten short, pointed denticles, median plates in four; a central row of VoL. XXIX] STOHLER: TWO NEW SPECIES OF GASTROPODS 431 Fig. 7. Same animal as shown in figure 6 withdrawn into its shell; note the black mantle rim outlining the operculum. simple, curved ridges separates the median plates; lateral plates heavy, den- ticles brownish; median plates weaker, denticles lighter brownish; central ridges weak, light ivory colored. MEASUREMENTS OF THE HOLOTYPE: Greatest height 149 mm.; maximum diameter 134 mm.; height of body whorl 64 mm.; maximum diameter of aper- ture 71 mm.; angle of divergence 62°; Operculum: greatest length 62.5 mm.; greatest width 47.8 mm.; maximum thickness 16.6 mm. PaRATYPES: In addition to the holotype, seven shells are hereby designated as paratypes I to VII; they include the specimen found by Dr. Fager on July 28, 1957 (paratype II), and five of those collected on November 19, 1957 (paratypes III to VII). Not included are the juvenile specimen (although some of its characters will be listed for the sake of completeness) and the still surviving animal as well as a dead shell found even earlier by Mr. Ray Ghelardi at the “Rockpile.”” Paratype I was the second specimen picked up. It was collected with the holotype by Cecil D. Jennings some time in De- cember, 1956, at what he stated was the same spot as that at which para- type II was found. 432 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. To facilitate comparison, the following table lists essential measurements and other observations of the paratypes and of the young animal, and ineludes the information already given concerning the holotype. TABLE 1 Comparative data for all the specimens of Astraea rupicollina, new species. considered in this study. x ~ Sy o = 8 Ss s te! Sn tal ES bo Sy ss] 3 3&8 S x= & = oa aS 2 a) oy to ™ os x Sie S85 eee Sa ees 23) a5 = BS S Corrugations on Silas TS Cae ES SES = © 8] Cs ER ER] x ~ o- base2 2.8/8 2 I & 2 2S Ed do ase 2-5 > 2 SS a iS SA iS Sslys RS So = S eS) 2 = 8 su R k Sw — or RF s Corrugations present — a) S oie ees sé <55 > on base of body oe a2 El Es | SS 5] Sw] Ss] go & whorl? SUE SS s|ESe| Bes] SS] Ses] SS S882) 33 S/SSS] SSS] 2S] SSS <8 eSS8/C 8] Remarks a 131% ol fi Gl |] Gl b 132 d|dj|d c 135 Gl || Gl |} Gl decollate d 137 1D) || al || cl slightly decollate e 131 1B) |}.1D) || 10) f 125 1D) || Gl |) Gl fq 131 d| dj - h 124 Gl || cl || Gl decollate i 122 D|djid decollate k 127% d|dj- decollate ] 135% —-|did m 125 djdjd much eroded n 119% d|Dj{d (0) 118 10) |/ 3D) 4) 10) slightly decollate Dp 115% D|djd q 115% 1D) |} Gl || Gl r 121% Gl |) cl |} cl s 108% ol] cl || 4 i 100% 10) || Gl |) te u 951% d|d|d 1. T =strong trace; t = trace; D = very distinct; d = distinct; —- = absent. 2. The axially aligned nodes of the whorls show a greater or lesser tendency to form secondary smaller nodes; the greatest tendency (i.e. the most distinctly subdivided nodes) are indicated with +++, while the complete absence of a subdivision is marked with —; logically ++, and + are intermediates between the two extremes. This column refers only to the first node above the carina, these nodes having a tendency to form what may be called a secondary carina. 3. The tendency toward forming subdivisions in addition to those forming the secondary carina is indicated Ly a + sign; the — sign indicates that the major nodes show no further subdivision. 434 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Tyre Locauity: “Rockpile,” approximately eight miles southeast of the southern tip of South Coronado Island, Lower California, Mexico; approxi- mately at 32° 17’ 36” North Latitude, 117° 09’ 30” West Longitude; in 70 feet of water; bottom: rocks, rockledges. The holotype will be deposited in the U. S. National Museum (No. 610, 331); paratypes will be deposited as follows: Stanford University collection (Stanford Univ. Paleo. type collection no. 8644)—paratype I; California Academy of Sciences in San Francisco (Calif. Acad. Sci. Dept. Geol. type col- lection no. 10471)—paratype II; Academy of Natural Sciences, Philadelphia —paratype III; Museum of Comparative Zoology, Harvard University, Cam- bridge, Massachusetts—paratype IV; Naturhistorisches Museum, Basel, Switzerland—paratype V; British Museum, London—paratype VI; San Diego Natural History Society Museum, San Diego—paratype VII. The species name is derived from the Latin rupes signifying rock and collis meaning hill, this being the nearest possible to a literal translation of the local name ‘“‘Rockpile” of the type locality. For purposes of comparison a group of 20 shells of Astraea undosa, col- lected on July 21, 1957, off La Jolla Point in 60 feet of water, were measured and the results entered in table 2. Corresponding measurements of the shells of Astraea petrothauma Berry are as follows: TABLE 3 Corresponding data for the shells of Astraea petrothauma Berry from Sacramento Reef (A) and from South Coronado Island (B) as well as those taken from the type figure (C) and the type description (D). = = = — PSs — ‘ Corrugations on 2 a) Seyeaee Sag |S a ES 8 base of bod Se} ad —— SS SS NS Spc ase of body os | os BS Ss ss SES s = 2 tS Rw Ss ~ =~ o 2. z SS [ESO ee] oS] eSvoes © & whorl S'S | oS 2 e= neq toe Seger Sw 38 we: bon ~—S . 5S Rod Bew Oo Sg 3 Bsr a SO re R k BSS|RSS| RSE | SSS] TS So] Gee La) a r top eroded top eroded 34.5 20.5 | 24.5 40.2 aE 22.2 | 25.6 39.3 | 49.3 1. This number is taken from Berry’s description of the holotype and attempts to convey that there are four more or less distinct corrugations plus a trace of a fifth one; the relative position of the various corrugations within the scheme shown in fig. 8 cannot be determined. 2. = means that both carinae are equally strong. To facilitate comparison of the three species in some particular respect, as well as to illustrate the variability within one or the other species in that same respect, a scheme was devised. It is noticed in Astraea undosa that the VoL. XXIX | STOHLER: TWO NEW SPECIES OF GASTROPODS 435 base of the body whorl may have up to six concentric corrugations. Our fig- ure 8 represents a diagram in which these six possible corrugations are num- bered from 1 to 6. C indicates the carina. Fig. 8. Diagrammatic representation of the six possible corrugations on the base of Astraea, numbered for convenience of reference. C =the carina. In the tables an attempt is made in column 6 to indicate the position of the corrugations in each specimen as well as to represent in some manner the relative development of each corrugation. If the corrugation is of ordinary or average development, it is assigned a “d” but if it is particularly strongly developed, that fact is indicated by a “D”; a “t” is meant to convey that there is a faint trace of a corrugation discernible, but it may be so faint that only tilting the shell back and forth in oblique light will reveal it; ‘‘T”’ on the other hand, indicates the presence of a readily recognizable trace. Placing the “t”’ between space 4 and 5 in table 1 was to convey the fact that here the trace is approximately equidistant between the other two corrugations. COMPARISON OF THE THREE LIVING SPECIES OF Astraea: Astraea rupi- collina differs from both A. wndosa (Wood) and A. petrothauma Berry by being distinetly higher spired, a much smaller angle of divergence, the rounder aperture, the marked convexity of the whorls, the rounded base of 436 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. the body whorl, the number and the arrangement of the corrugations on the base of the body whorl, as well as in details of the sculpture. Many shells of A. wndosa examined showed that apparently the nodes, constituting the carina, are laid down first as completely hollow outpocket- ings of the body whorl; if one examines the inside of the body whorl, it becomes apparent that the animal seems to fill in these empty pockets with nacreous material because, the further inward one proceeds, the less deep are the outpocketings and about one and a half inches from the aperture, the inside of the body whorl appears smooth and devoid of pockets. In our specimens of A. petrothawma both carinae show this condition but not one of our A. rupicollina shells has a hollow carina, except the young animal, where the lower carina shows an internal fairly shallow and inwardly rapidly diminishing groove; it does not compare at all to the depth of the carinae in the other two species. The living animal of Astraea rupicollina differs from that of A. wndosa in several features: in the adult of A. wndosa only the first epipodial tentacle is observed and at that it is difficult to see (fig. 9), while in A. ruprcollina all four tentacles are very distinct and rather eye-catching because of their con- trasting color; the color of the sides of the foot of A. rupicollina is essentially ccm Fig. 9. Living animal of Astraea undosa (Wood), showing the mottled pattern of the foot; note the single epipodial tentacle. Vou. XXIX] STOHLER: TWO NEW SPECIES OF GASTROPODS 437 two-toned, separated into two more or less distinct parallel areas, while the sides of the foot of A. wndosa are more or less uniformly mottled and of a much lighter general color, although the dark parts of the mottling are darker (M. and P. pl. 56, A 1) than any of the colors on A. rupicollina; the mantle fold in A. undosa is not outlined in black around its entire periphery, but is limited to the area outside of the aperture, nor is the yellow tinge present. No color comparison with A. petrothauma was possible as no living speci- mens were available for study. There are some differences also in the operecula: The four ribs of the operculum of Astraea undosa as well as most of the margin itself are covered with numerous, in part sharply pointed rugosities; they are lacking on the margin opposite the nucleus and on the distal three-fourths of the second rib (numbering the ribs from the basal side up when the opereulum is with- drawn into the aperture). In A. petrothauma (both specimens, from Sacra- mento Reef and from South Coronado Island) the first rib is almost entirely smooth, the second has relatively delicate rugosities for about three-fourths of its length, and the third rib has more pronounced rugosities for about seven-eighths of its length. The nucleus of the opereulum of A. rupicollina appears less spirally recurved than in A. petrothauma; of the three ribs the first has few, the second slightly more rugosities, all of which are almost obsolete; on the third rib the sharply pointed pustules extend only for about two-thirds of the length. The margins in both these latter two species are entirely smooth. There are also some differences, though perhaps slight, in the coloration of the periostracum of the opereula. In A. wndosa the inere- mental ridges are coarser, while the longitudinal ridges are so faint as to be almost absent (see fig. 5b); the overall effect is much more uniformly dark; in A. petrothauma the incremental ridges are distinct, but much finer than in either of the two other species, and the longitudinal ridges are not at all discernible; in A. rupicollina the longitudinal ridges are more distinct than in A. undosa and are present also in the nepionie portion. The question arises whether the operculum illustrated by Berry and described as the paratype (J. c., pl. 25, fig. 3) is actually the operculum of the adult Astraea petrothauma or, possibly, that of the present species. If the latter were the case, then Berry’s name as applied to the operculum would fall into the synonymy of A. rupicollina, and the name, A. petrothawma would then apply only to the shell described in the same paper. However, I feel that the material at hand is not sufficiently large to warrant a decision at this time; it is hoped that additional specimens of Berry’s species, including living animals, may be collected at some future date. In concluding the discussion of this portion of the present paper, it may be well to consider two other points. It seems, first of all, rather strange that 438 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. such a large and magnificent animal as Astraea rupicollina has not turned up in dredgings long before this. The nature of the type locality may account for this: beeause of the large rocks and ledges a dredge, if one were used, might well skim above and over the shells. On the other hand, the manner of the discovery of this species points up the great value of the relatively new research technique employing SCUBA (self contained underwater breathing apparatus). Many exciting discoveries may be expected as this new technique is refined and its application expanded. Secondly, the fact that A. rupicollina so far has not been found anywhere else than on the hilltop of the “Rockpile,” not even in neighboring areas which were scanned for it during the search in July, 1957, coupled with the fact that A. wndosa is indeed a highly variable species, may possibly lead, eventually, to the conclusion that we are dealing with an endemic geographical variant of A. undosa. It might be further considered possible that A. petrothauma is still another subspecies in the wn- dosa-Rassenkreis. Until the areas between the “Rockpile” and South Coro- nado Island in one direction, and between the “Rockpile” and Sacramento Reef in the other, are carefully and minutely explored, it seems advisable to consider all three forms as good species. And since no specimens of A. undosa were obtained at the ‘“Rockpile” or anywhere nearer to it than at the south tip of South Coronado Island, approximately 8 miles to the northwest, coupled with the fact that no tendency toward the A. rupicollina or A. petro- thauma shape could be observed in the many examples of A. undosa examined, the idea that we are indeed dealing with a heretofore unknown species seems to gain plausibility. The fact that a living specimen of Astraea petrothawma Berry was found very recently at South Coronado Island, does not, it would seem, detract from the validity of this conclusion; it adds, however, to the problem of dis- tribution of Berry’s species. Originally it was not deemed worthy to note that the specimen of A. petrothauma from Sacramento Reef was taken in less than 30 feet of water. Since, however, the second specimen at South Coronado Island was also taken in relatively shallow water (between 10 and 50 feet, according to Mr. Limbaugh, although the exact depth where the specimen in question was picked up, is unknown), and since, on the other hand, all of the known specimens of A. rupicollina come from depths in excess of 70 feet, it now seems proper to record these facts. CoMPARISON WITH Astraea gradata GRANT & GALE: In comparing Astraea rupicollina with the fossil species from the middle Pliocene, described in 1931 by Grant and Gale as A. gradata, several points of difference are readily noticeable. The material used for this purpose con- sisted, in addition to the excellent illustrations of the type material in Grant & Gale, plate 31, of two topotypical specimens from the collection of the VoL. XXIX] STOHLER: TWO NEW SPECIES OF GASTROPODS 439 Museum of Paleontology at the University of California, kindly made avail- able to me by Dr. J. W. Durham. In Astraea gradata the upper of the two parallel carinae is much less fully developed, especially on the early whorls, than in the new species. The nodes on the whorls are raised more and turn in at almost a right angle toward the suture, thus giving the fossil species that more or less tabulated aspect which led the authors to choose the name; in A. rupicollina the nodes appear more or less parallel in their entire length to the convexity of the whorl. There are three basal corrugations in A. gradata, equally distributed over the entire basal area, not leaving room for a possible undeveloped fourth or even fifth corrugation, a condition clearly present in A. rupicollina (see the comparison above with A. wndosa). Even more striking are the differences in the opereula. That of A. gradata is sub-circular and seems to have only two ribs, although there is a broad, very flat, low protuberance, possibly a rudiment of the (for the subgenus Uvanilla) typical third, basal rib. While the middle rib of A. rupicollina is thin, high, and overhanging, that of A. gradata is flat, shallow, and broad. The area between the two basal ribs in A. gradata appears to be covered with fairly numerous, fine pustules, whereas the corresponding area in A. rupicollina is polished. Finally, the nuclear whorl of the operculum is much less recurved in A. gradata than it is in the new species. In spite of the many differences between the two species, it seems possible that Astraea rupicollina may prove to be a more or less direct linear descend- ant of A. gradata. Much more fossil material of the latter and some very young stages of the former species will be needed before this question can be answered satisfactorily. Macrarene coronadensis Stohler, new species. As mentioned above, in the search for the new astraea, a number of sta- tions around the Coronado Islands were closely examined by Dr. Fager and his associates during July, 1957. Only a few hundred feet seaward from the North Island they found six specimens of a species belonging to the genus Macrarene (Hertlein and Strong, 1951). The same species was taken on three different dives one week later at the “Rockpile”’; on November 19, 1957, Dr. Fager picked up one living and one dead specimen, again at the “Rock- pile.””’ Depths of collections varied from 150 feet to 70 feet. The shells did not agree with the three other species of the genus, found in California waters, i. e., Macrarene pacis (Dall), M. californica (Dall), and M. farallonensis (A. G. Smith). The rich material in the collections of the San Diego Society of Natural History and of the California Academy of 440 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47H SER. Sciences was minutely compared before it was decided to consider this a new species. For the description it became necessary to compile the characters from three shells (one of which is designated the holotype and the other two para- type I and II respectively) because all shells are more or less completely covered with enecrusting organisms and because it proved impossible to re- move this encrustation without possibly damaging the finer sculpture of the shell, at least in part. Fig. 10. Macrarene coronadensis, new species. Holotype and paratypes I and II. Top row: apical view of the three specimens; middle row: the same three specimens seen from the umbilicus; bottom row: lateral view. The first picture in each row is of the holotype; the second of paratype I and the third of paratype II. Norr: On the holotype may be noted a peculiar, extremely hard encrustation surmounting the apex and giving it a somewhat higher aspect than it is in reality. Photographs by Victor Duran. VoL. XXIX] STOHLER: TWO NEW SPECIES OF GASTROPODS 441 SHELL: Large for the genus, heavy, tabulated, with three and a half to four whorls, nacreous-white; periphery of whorls angular with 9 to 12 pro- jections on the body whorl and the preceding whorl; tabulated area above the projections smooth, unsculptured except for very faint undulations coin- ciding with the projections. Whorls rapidly enlarging with distinct sutures; aperture circular within, shiny, pearly; outer margin of aperture flaring, with five corners, best described as a square to the base of which is fused an isosceles triangle. The whorls are angular in outline; this outline is pro- duced by the relatively sharply pointed, subequally spaced, short projections of the periphery; the projections are connected with fairly distinct ridges producing the effect of a continuous carina. Parallel to the peripheral carina and below it is a second carina, equally strong and consisting of the same elements, 7. €., projections and connecting ridges. The projections of the two carinae are connected with each other by oblique, somewhat folded-over ridges which divide the otherwise smooth lateral area of the whorl into sub- equal parallelograms. Below the lower carina the body whorl slants toward a narrow, rounded base. In addition to the faint ridge-lke oblique swellings corresponding to the projections above, there is on this portion of the body whorl a pair of spiral cords, very closely situated to each other and about equidistant between the lower carina and the base of the whorl. The umbilicus is very wide, deep, spiral with a distinct keel spiraling down and ending on the inner base of the isosceles triangle. Whorl ending obliquely (see fig. 10). OpercuLUM: Horny, with numerous, fine, concentric rings of ridges (this is very reminiscent of the operculum of Norrisia norrist (Sowerby), except that the rings are so fine that they are discernible only under magni- fication) ; there is a fringe of fine strands of horny substance around the periphery; the shallowly coneave operculum seems to fit “airtight” over the inner aperture. ANIMAL: Sides of the foot brown; foot narrow, fairly long, about two and a half times as long as broad; tentacles filiform. HOLOTYPE MEASUREMENTS: Maximum height 18.6 mm.; max. diam. 22.5 mm.; body whorl height 11.4 mm.; diam. of aperture 6.1 mm.; angle of diver- gence 122°. MEASUREMENTS OF PARATYPES I AND II: I. Max. height 18.7 mm.; max. diam. 22.4 mm.; height of body whorl 12.0 mm.; diam. of aperture 6.3 mm.; angle of div. 128°. II. Max. height 15.8 mm.; max. diam. 22.7 mm.; height of body whorl 11.6 mm.; diam. of aperture 6.5 mm.; angle of div. 14014°. TYPE LOCALITY: Seaward side of North Island of the Coronado Islands, Lower California, a few hundred feet from shore in 150 feet; bottom: large rocks, ledges, sand pockets. This locality is at about 32° 27’ North Latitude and 117° 11’ 45” West Longitude. 442 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. In table 4 are recorded the measurements of all 25 specimens collected alive in the two areas deseribed above (for comparison the measurements of the holotype and paratypes I and II are also entered in the table). TABLE 4 Comparative data for all 25 specimens of Macrarene coronadensis, new species. s = = ES we iS 3 > S = = | = Seal) sas CES Sol hacia Baars es | sal = = 25 oS] SSS 3&8 aa 3 A Seeey | ee as B22] ses at >3 > Sars aS ma} Ses| SS SRS S Sz ss SV ]_Ssss r 2 S) ° SS; Ss] es iQse| <8 | ss] & Reman Holotype : W222 Paratype I : 128° II d 140%° III } 118° ? heavily encrusted IV 5 106%° ? heavily encrusted V : 110° ? nepionic whorls encrusted Hypotype I 3 109%° II ; 1191%° heavily encrusted III : 23 IV 4 123% V : 1261%4° J ? heavily encrusted VI : 130° 3% VII : 130° VIII : 130° 4 IX : 126° ? heavily encrusted xe : 1262 4 XI : 12314° ? heavily encrusted XII 124° 3Y XIII i 134° 3% XIV : 136° 3 XV : 134%4° 4 XVI d 138° 2 heavily encrusted XVII : 1321%° 3Y XVIII , 140%° ? heavily encrusted XIX d 121%° 3 The specimens collected at the ‘“Rockpile” are designated as hypotypes T to: XLX DISPOSITION OF TYPES: The holotype and paratypes I and II will be de- posited in the U. S. National Museum in Washington, where they will be assigned the numbers 610,332 (holotype) and 610,333 (paratypes). Para- type III will be deposited at Stanford University (Stanford Univ. Paleo. type coll. no. 8645); paratype IV will go to the California Academy of Sciences (Calif. Acad. Sci. Dept. Geol. type coll. no. 10506); paratype V Vou. XXIX]} STOHLER: TWO NEW SPECIES OF GASTROPODS 443 will be sent to the British Museum in London. Of the many hypotypes, one each will be deposited with the Academy of Natural Sciences, Philadelphia; Museum of Comparative Zoology, Harvard University; Naturhistorisches Museum Basel, Switzerland; and the San Diego Natural History Society Museum in San Diego. The remaining hypotypes will be retained, at least for the time being, in the collection of the Department of Zoology at the University of California in Berkeley. CoMPARISON: Macrarene coronadensis appears to be larger than any of the three other species of this genus known from California waters. It has more projections on the periphery of the body whorl than either of the two Dall species (I. pacis and M. californica) and generally fewer than the third species, M. farallonensis. It bears the greatest affinity to M. pacis but is easily distinguished from it by the possession of the second row of projections below the periphery. In both species described by Dall the projections at the periphery are curved upward, while in J/. coronadensis they project horizontally. From M. farallonensis the new species is easily distinguished by the absence of the pronounced axial sculpture on the basal portion of the body whorl as well as by the smoothness of the relatively level upper surface of the body whorl. Because of these marked differences a rather simple key may serve to separate the four species now known. KEY TO THE CALIFORNIA SPECIES OF MACRARENE 1. Projections at the periphery of the body whorl six, blunt, rounded...........-.-..-----.- nce se StS OSEP ey Se ee RIE woe M. californica (Dall) — Projections at the periphery of the body whorl more than six... 2 2. Only one row of projections on the body whorl, not exceeding eight in number_____... en eee te RE ee ae cu schon mecca anne Bea ee M. pacis (Dall) — Body whorl with more than eight projections or with two parallel rows of pro- EJ CE] ONS os eee te et a a Ea Sa atc ee ES, Dee 3 3. Body whorl with two distinct, equally strong parallel rows of projections; 7 to 12 projections per row (usually 9 or 10); projections not pointing upward; upper suntaceioh whorls rather smooth = M. coronadensis Stohler — Body whorl sculptured with strong axial oblique ribs extending from suture to base ee ee i IO AS 8 eS he ere ee M. farallonensis (A. G. Smith) Remarks: There has been much discussion regarding the correct generic name for this group of shells; Smith in 1952 gave a brief, yet excellent sum- mary of this problem. However, I feel that the generic name proposed by Hertlein and Strong in 1951 really beautifully and simply solves the prob- lem and I have adopted that name for those reasons. It seems fitting to close this paper with an expression of appreciation for the generous cooperation shown by so many people, without which these two species might have remained unknown still longer. The roles of Dr. 444 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47TH SER. E. W. Fager and R. J. Ghelardi, as well as those of Mr. and Mrs. C. D. Jen- nings and Mrs. Alan H. Wolfson have been mentioned. In addition to these contributions, credit belongs also to C. M. Boyd, T. K. Chamberlain, W. D. Clarke, H. L. Scotten, and J. R. Stewart, graduate students at Seripps In- stitution of Oceanography at La Jolla for spending much time in diving and helping in the search; Mr. C. Limbaugh, the marine diving specialist of the same institution also participated in the collecting; without the good-na- tured accommodations of Mr. D. A. Noland, the skipper of the buoy boat used, the search would probably not have been successful. Drs. A. M. Keen of Stanford University and Leo G. Hertlein of the California Academy of Sciences generously made the collections under their curatorship available to me. Dr. Cadet Hand of the Department of Zoology, University of Cali- fornia, kindly read the manuscript. The figures owe their excellence to the great skill of the staff artist, Mrs. Emily Reid. BIBLIOGRAPHY Berry, 8S. STILLMAN 1940. New Mollusca from the Pleistocene of San Pedro, California—I. Bulletin of American Paleontology, 25:149-164. DALL, WILLIAM HEALEY 1908. The Mollusca and the Brachiopoda. In: Reports on the Scientific Results of the Expedition to the Eastern Tropical Pacific, in charge of Alexander Agassiz etc. Bulletin of the Museum of Comparative Zoology, Harvard, 43 : 205-487. GRANT, U.S., IV, and Hoyt RopNEY GALE 1931. Catalogue of the Marine Pliocene and Pleistocene Mollusca of California and Adjacent Regions. Memoirs of the San Diego Society of Natural His- tory, 1:1036 pp., 15 text figs. and 32 plates. HERTLEIN, LEO GEORGE, and A. M. StronG 1951. Mollusks from the West Coast of Mexico and Central America. Part X. in: Eastern Pacific Expeditions of the New York Zoological Society. Zoologica, 36:67—-120. KEEN, A. MyRA 1956. Nomenclatural problems in the Archaeogastropoda. Annual Report of the American Malacological Union, Bulletin 23: 6-7. Smitru, ALLYN G. 1952. Shells from the bird guano of southeast Farallon Island, California, with description of a new species of Liotia. Proceedings of the California Academy of Sciences, 4th ser., 27:383-387. Srronea, A. M. 1934. West American species of the Genus Liotia. Transactions of the San Diego Society of Natural History, 7:429—452. PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES FOURTH SERIES Vol. XXIX, No. 12, pp. 445-463 February 27, 1959 REPORT ON A COLLECTION OF REPTILES FROM AFGHANISTAN BY ALAN E. LEVITON Department of Herpetology California Academy of Sciences During the months of March through September, 1950, Mr. John Gas- peretti, on a petroleum exploration survey in Afghanistan, made a small collection of reptiles which he subsequently presented to the California Academy of Sciences. These reptiles, 65 specimens representing 19 species, were collected at the town of Chah-i-Angir’, in the Dasht-i-Margo Desert, approximately 100 miles southwest of Khanadar, Afghanistan. There are no new forms in this collection. Several species, however, are for the first time definitely recorded from that country. Following the systematic notes on the collection, I have summarized what is currently known of the affinities of the herpetofauna of Afghanistan and have prepared a checklist of the reptiles and amphibians presently known from that country. In regard to the specimens at hand, Mr. Gasperetti has kindly supplied the following information about the region in which he made the collection : 1 The locality Chah-i-Angir has been found on the Girishk quadrangle, no. 34-A, of the 1940 edition of Army Map Service map 103127, 1942, International number H-41 E. It has been determined to be 64° 26’ 55” E. Long., and 31° 40’ 53” N. Lat., approximately 16.5 miles north of the confluence of the Helmand and Arghan- dab rivers, southern Afghanistan. [445] 446 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. “Chah-i-Angir is at the confluence of the Helmand and Arghandab rivers, about 100 miles southwest of Khandahar, in the Dasht-i-Margo Desert. The Dasht-i-Margo is a high desert... . the elevation is about 1,000 feet above sea level. There are large sandy areas and intermittent alluvial gravel plains with sparse, small vegetation. Gazelle and lesser bustard abound. The weather in this area is severe, ranging from 120°F. summer tempera- ture to near 0°F., with snow, in winter. Daily extremes are probably 50 degrees or more, as for example 50°F. to 120°F. in summer and 30°F. to 70°F. or 80°F. in winter. The dry season extends from April through September.”’ SYSTEMATIC NOTES Class REPTILIA Order SAURIA Family GEKKONIDAE Agamura persica (Duméril) Gymnodactylus persicus DUMERIL, 1856, Arch. Mus. Hist. Nat., Paris, vol. 8, p. 481 (type loc.: Iran). Agamura persica BLANFORD, 1876, Zool. E. Persia, p. 358, pl. 23, fig. 4a—b. BOULENGER, 1889, Trans. Linnean Soc. London, ser. 2, vol. 5, p. 95, pl. 9, fig. 2. SMirn, 1935, Fauna Brit. India, vol. 2, p. 61, fig. 19. MATERIAL EXAMINED (2): CAS 84690-84691. Remarks: The identification of these specimens is somewhat in doubt. Agamura persica and Gymnodactylus stoliczkai are very similar in appear- ance except for the structure of their tails, which differs very strikingly. The tails, however, are missing from both of our specimens, only a few millimeters of stump remaining. The stump is not broadened as in G. stoliczkai but rather normal and cylindrical. On the other hand, there are only 10-11 upper labials, fewer than recorded for A. persica, but within the limits for G. stoliczkai. Since the structure of the tail is certainly the more important of these two taxonomic characters, I have assigned the specimens to the former species. The two specimens at hand are juveniles, measuring 22.5 mm. and 28.0 mm. in standard (snout—vent) length. In both individuals there are 10-11 upper labials; the seales of the dorsum are heterogenous, with somewhat en- larged, rounded and slightly keeled tubercles scattered among the small VoL. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 447 juxtaposed granules; the ventral seales are slightly larger than those of the dorsum and are imbricate; and there is an indistinct pattern of darker eross-bands on the otherwise pale grayish background of the dorsum. Alsophylax tuberculatus (Blanford) Bunopus tuberculatus BLANForD, 1874, Ann. Mag. Nat. Hist., ser. 4, vol. 13, p. 454 (type loc.: Baluchistan) ; 1876, Zool. E. Persia, vol. 2, p. 348, pl. 22, fig. 4. Alsophylax tuberculatus BOULENGER, 1885, Cat. Lizards Brit. Mus., vol. 1, p. 20. Sm1rH, 1935, Fauna Brit. India, vol. 2, p. 36, fig. 14. MATERIAL EXAMINED (1): CAS, unnumbered specimen. Remarks: This individual was taken from the stomach of our specimen of Lytorhynchus ridgewayt. The head had been largely digested. Other- wise the body is in good condition and the characteristics of the toes and tubereulation of the body are readily discernible. The specimen lacks pre- anal pores and presumably is a female. Teratoscincus cf. T. bedriagai Nikolsky Teratoscincus bedriagai NiKo.Lsky, 1899, Ann. Mus. Zool. Acad. Imp. Sci. St. Peters- bourg, 1899, p. 146 (type loc.: eastern Iran). MATERIAL EXAMINED (1): CAS 84689. Remarks: The identification of this specimen is only tentative. This conditional identification results from the fact that 7. bedriagai may be conspecific with 7. prezewalski, known from central China and eastern Turkestan, from which species 7. bedriagai has been distinguished by the single character of relative size of ventral versus dorsal scales. According to Bedriaga (1905, pp. 159-162) the ventral scales are equal to or slightly larger than the dorsals in 7. prezewalski, and are slightly smaller than the dorsals in T. bedriagar. Both species are readily distinguished from other members of T'eratoscancus by a combination of two conspicuous char- acters: 1) the enlarged seales on the dorsum do not extend forward beyond the shoulder, and 2) there usually are less than 50 scales around the middle of the body. In the specimen at hand there are 52 seale rows around the middle of the body, and the enlarged and imbrieate cycloid scales of the dorsum do not extend forward beyond the anterior border of the shoulders. The ventral seales are either subequal to or slightly smaller than the dorsals. There are 12 granules in a longitudinal row between the anterior corner of the eye and the nasal shields, 80 granules in a longitudinal row extending from the snout to the first of the enlarged dorsal cyeloid scales, and 55 granules 448 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 47TH SER. in a median longitudinal series extending from the mental to the first en- larged ventral scales. Nine upper and nine lower labials are present. The nostril is bordered by the rostral and four cireumnasal shields. Measurements (in mm.): Standard length 46; tail length 25. Teratoscincus scincus (Schlegel). Stenodactylus scincus SCHLEGEL, 1858, Handl. Dierk., vol. 2, p. 16 (type loc.: Ili River, Turkestan). Teratoscincus keyserlingii STRAvUCH, 1863, Bull. Acad. Sci. St. Petersbourg, vol. 6, p. 280 (type loc.: Seri-Tschah, Iran). Teratoscincus scincus BOULENGER, 1885, Cat. Lizards Brit. Mus., vol. 1, p. 12, pl. 2, fig. 3; 1889, Trans. Linnean Soc. London, ser. 2, vol. 5, p. 94, pl. 8, fig. 1. SmiruH, 1935, Fauna Brit. India, vol. 2, p. 30, fig. 12. Teratoscincus zarudnyi NIKOLSKyY, 1897, Ann. Mus. Zool. Acad. Imp. Sci. St. Peters- bourg, 1897, p. 309, pl. 18, fig. 1 (type loc.: eastern Iran). Teratoscincus roborowskii BepRIAGA, 1905, Ann. Mus. Zool. Acad. Imp. Sci. St. Peters- bourg, vol. 10, p. 159 (type loc.: Oase Ssatschsheu [= Oasis of Sachow, Kansu, fide Pope, 1935, Rept. China, p. 458]). MATERIAL EXAMINED (2): CAS 84648-84649. ReMaArRKS: Of seven nominal species presently assigned to the genus Teratoscincus, three are regarded as conspecific with 7’. scincus. Terentjev and Chernoy (1949, p. 128) have suggested that differences in scutellation and color pattern may permit recognition of 7. keyserlingi and T. ro- borowski as distinct subspecies of 7’. scincus. Our specimens agree with the description of 7. scincus given by Smith (1935, p. 80). In CAS 84649, an adult male, there are 31 scales around the body; the enlarged dorsal scales, 49 in a longitudinal series counted from the first enlarged seale on the occiput to the first of the large caudal scales, extend in a band six seale-rows wide from the shoulder to the occiput. There are 35 granules between the eyes (about 45 in CAS 84648), and 17 gran- ules between the eye and nasal shields. The nostril is bordered by a rostral and four circumnasal shields, and there are 11 upper and 11 lower labials. A few small blackish spots are arranged more or less longitudinally on either side of the middorsal line. There is a seattering of reddish spots in the shoulder region, and two diagonal lines on each side of the neck extend from behind the eye to the arm. Otherwise, the ground color is pale gray. The skin, which among species of Teratoscincus is extremely delicate, has been badly damaged in CAS 84648 and body scale patterns cannot be determined. Measurements (in mm.) are as follows: CAS 84648 CAS 84649 Standard deneth: ssa... -.208 90.0 98.5 Waris Were lates tae eeeperci ay ccs cece - ve 31.0 33.0 Vou. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 449 Family AGAMIDAE Agama agilis Olivier. Agama agilis OvIvier, 1807, Voy. Emp. Otho., vol. 4, p. 394, pl. 24 (in atlas), fig. 2 (type loc.: vicinity of Baghdad, Iraq). Smiru, 1935, Fauna Brit. India, vol. 2, p. 221. Agama isolepis BOULENGER, 1885, Cat. Lizards Brit. Mus., vol. 1, p. 342; 1889, Trans. Linnean Soc. London, ser. 2, vol. 5, p. 96, pl. 10, figs. 1-3. MATERIAL EXAMINED (6): CAS 84642-84647. REMARKS: Three of this series of six specimens are females. These are all somewhat lighter on the dorsum than the males; all have black stripes on the throat, but these are very light. The females lack the distinctive black ventrolateral patches which are present in the males. In two of the females, CAS 84645 and 84646, the anterior portions of the head are light gray, which stands in sharp contrast with the much darker grayish black of the dorsum. The throats of all the males are striped grayish black, the dark color tending to suffuse over the entire throat surface. The black ventrolateral patches do not meet midventrally, but are separated by a narrow grayish space. In CAS 84647, the upper labials, nasals, rostrals, and the anterior half of the throat are light gray; in CAS 84644, the upper and lower labials similarly are hght gray. Eggs were found in all females: CAS 84642 contained five eggs, the largest, measuring one centimeter in diameter, appeared ready for expul- sion; in CAS 84645, about 15 eggs were counted in a grape-like cluster at the ovary, the largest measuring only two millimeters in diameter. The testicular tissues in all male specimens appeared to be greatly enlarged. Tenebrionid beetles were found in the stomachs of several individuals. Measurements (in mm.) and counts are summarized in table I. TABLE I[ Measurements (in mm.) and counts for specimens of Agama agilis CAS Standard Tail Scale Upper Lower Preanal number Sex length length rows labials labials pores 84642 2 90 136 74 16 18 0 84643 3 94 147 76 19 if 21 84644 3 ae 130 76 20 i7/ 1s 84645 Q 81 IIL 7 76 17 16 0 84646 je) 86 117 75 19/21 18 4/7 84647 3g 92 129 72 19 16 25 450 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. Phrynocephalus ornatus Boulenger. Phrynocephalus ornatus BOULENGER, 1887, Cat. Lizards Brit. Mus., vol. 3, p. 496 (type loc.: between Nushki and Helmand River, Afghan—Baluchistan frontier); 1889, Trans Linnean Soc. London, ser. 2, vol. 5, p. 97, pl. 8, fig. 3. Smiru, 1935, Fauna Brit. India, p. 232. MATERIAL EXAMINED (7): CAS 84650-84653, 84660-84662. Remarks: The four females and three males of this series agree closely with Smith’s description (1935, p. 232) of the species. Ovarian eggs were found in each of the females. MEASUREMENTS (range in mm.) for this species of specimens are as follows: Males Females standard leneth .......... 38.7-39.5 37.0-41.5 Maal engines: ci. socks iste 50.0—55.0 49.0-54.5 Phrynocephalus scutellatus (Olivier). Agama scutellata OLIVIER, 1804, Voy. Emp. Otho., vol. 3, p.110 (type loc.: near Ispahan, Iran); 1807, ibid., pl. 42 (in atlas), fig. 1. Phrynocephalus olivieri DUMERIL and BrIspron, 1837, Erp. Gen., vol. 4, p. 517 (based upon Olivier’s specimen). BOULENGER, 1889, Trans. Linnean Soc. London, ser. 2, vol. 5, p. 96, pl. 8, fig. 2. Phrynocephalus scutellatus Smirn, 1935, Fauna Brit. India, vol. 2, p. 229. MATERIAL EXAMINED (12): CAS 84654-84659, 84663-84668. RemMaArRKS: Eges were found in the oviducts of six of the seven females, and the testicular tissues were enlarged in all male specimens. Several ants were found in the stomach of one individual. The measurements (in mm.) for the sample are as follows: Male Female Standard eneth: s.6<.0% 4% 42.5-48.0 40.5-50.5 Merl Memo Pai teee 2 aie Piece aoa-ae ere 57.5-74.0 56.5-68.0 Family SCINCIDAE Ophiomorus tridactylus (Blyth). Sphenocephalus tridactylus Brytru, 1855, Jour. Asiatic Soc. Bengal, vol. 22, p. 654 (type loc.: Afghanistan). Sphenoscincus tridactylus Perers, 1875, Monatsb. Berl. Akad., 1875, p. 553, figs. 6-12. Ophiomorus tridactylus BOULENGER, 1887, Cat. Lizards Brit. Mus., vol. 3, p. 394. Smiru, 1935, Fauna Brit. India, vol. 2, p. 346, fig. 78. Vor. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 451 MATERIAL EXAMINED (7): CAS 84669-84675. Remarks: All specimens at hand are uniform pale cream above with a narrow dorsolateral dark brown stripe which originates at the nostril, passes through the eye, and extends onto the tail. In several specimens there is a dense spotting of dark brown on the posterior portion of the tail. There is no variation in head scutellation with the exception of CAS 84674 in which there are five rather than six upper labials, and the large fourth shield borders the eye; the loreal on the right side is divided into two small shields. Enlarged ovarian eggs were found in each of the two females. Sand grains and fragmentary insect remains were found in the stomachs of two specimens. The measurements for the sample are as follows: Male Female DSianGards-lemeph cs ..5/. seis = sea es 72-82 86-90 Teele emit hne hie ton. clea gee ieva Siete wees 56-65 66 Family LACERTIDAE Acanthodactylus cantoris cantoris Giinther. Acanthodactylus cantoris GUNTHER, 1864, Rept. Brit. India, p. 73 (type loc.: Ramnagar, Punjab, India). BLANrForD, 1876, Zool. E. Persia, p. 381, pl. 26, fig. 3. Acanthodactylus cantoris cantoris SmirH, 1935, Fauna Brit. India, vol. 2, p. 371. MATERIAL EXAMINED (6): CAS 84676, 84678-84679, 84681-84683. REMARKS: This appears to constitute the second record for this species from Afghanistan. Heretofore, the species has been recorded from eastern Iran, Baluchistan, and the North West Frontier Province, and once from southern Afghanistan (Boulenger, 1889, p. 101). Four of the six specimens in this series are females. In three of these, eggs were found in the oviduct, and in one the eggs were clumped togther in the ovary. A single egg in the oviduct of CAS 84682, an individual 63 mm. in standard length, measured 14 mm. along its long axis. Insect remains were found in the stomachs of several specimens. How- ever, in one individual, CAS 84683, the largest male of the series, the par- tially digested remains of a skink, Ophiomorus tridactylus, were found. As indicated by the data in table II, the difference between the sexes in the number of scale rows on the dorsum and in the number of femoral pores suggests these characters are subject to sexual dimorphism. There are no marked differences in coloration or in proportional measurements between the sexes, but the males seem to reach a slightly larger size. 452 CALIFORNIA ACADEMY OF SCIENCES [ Proc, 4TH SER. TABLE II Summary of measurements (in mm.) and counts for the sample of Acanthodactylus cantoris cantoris Character Male Female Morsal’scallerrOws)ess eres 38-41 33-37 Ventral scale rows= ee 13-14 12-14 Hemoralipores se See eee 21—23 18-20 Standand lene hee eee enna 66.0—70.5 63.0-66.5 Daal, Jen et hy ee -ee oe eee er eee 130 96-123 Eremias fasciata Blanford. Eremias fasciata BLANFOoRD, 1874, Ann. Mag. Nat. Hist., ser. 4, vol. 14, p. 32 (type loc.: Saidabad, southwest of Karman, Iran); 1876, Zool. E. Persia, p. 374, pl. 25, fig. 3. SMITH, 1935, Fauna Brit. India, vol. 2, p. 386. MATERIAL EXAMINED (5): CAS 84684-84688. REMARKS: Smith (1935, p. 386) records this species from southern Af- ghanistan but gives no exact locality. Boulenger (1889, p. 100) reeords a single specimen from Helmand. The five specimens reported on here, four males and one female, agree closely with Smith’s description of the species in seutellation, but differ in color pattern. According to Smith there are five to nine stripes on the neck which reduce to five to seven on the back. In our specimens there are nine to eleven stripes on the neck which extend, without reduction in number, onto the back. Measurements (in mm.) and counts for this series of specimens are summarized in table ITI. TABLE III Measurements (in mm.) and counts for specimens of Eremias fasciatus CAS Scale rows Upper Femoral Standard Tail number Sex Dorsal Ventral labials pores length length 84684 3 50 14 aly 17/19 58 99 84685 a 49 IBF; 12/13 19/18 62 — 84686 3 52 16 I Alit 17/15 47 _— 84687 3} 51 15 11/11 18/18 61 91 84688 fe) 46 15 12/12 16/18 42 83 Eremias guttulata watsonana Stoliczka. Eremias (Mesalina) watsonana SrTouiczKa, 1872, Proc. Asiatic Soc. Bengal, 1872, p. 86 (type loc.: between Karachi and Sukkur, Sind). Eremias guttulata watsonana SMITH, 1935, Fauna Brit. India, vol. 2, p. 389. Vou. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 453 MATERIAL EXAMINED (1): CAS 84680. Remarks: Previously recorded from the Afghan—Baluch border by Aleock and Finn (1896, p. 558) and from Afghanistan proper by Boulenger (1889, p. 99) and Smith (1935, p. 390; 1940, p. 384) this subspecies has also been recorded from Sind and the North West Frontier Province to the east and from Iran on the west. Lantz’s record (1918, p. 15) of EF. gut- tulata from the River Tajan, Transeaspia, probably belongs here. Our specimen is a gravid female measuring 49 mm. in standard length and 89 mm. in tail length. A single egg was found in the left oviduct. There are 44 longitudinal rows of dorsal scales and ten longitudinal rows of ventrals. No preanal or femoral pores are present. Eremias velox persica Blanford. Eremias persica BLANFORD, 1874, Ann. Mag. Nat. Hist., ser. 4, vol. 14, p. 370 (type loc.: near Ispahan, Iran) ; 1876, Zool. E. Persia, p. 370, pl. 26, fig. 1. Eremias velox persica BOULENGER, 1921, Monog. Lacertidae, vol. 2, p. 312. Smirn, 1935, Fauna Brit. India, vol. 2, p. 383. MATERIAL EXAMINED (1): CAS 84677. REMARKS: This single male specimen measures 87 mm. in standard length and 172.5 mm. in tail length. Its seale counts fall within the range for the species given by Smith (1935, pp. 383-385). Family VARANIDAE Varanus (Psammosaurus) griseus (Daudin). Tupinambus griseus DAuDIN, 1803, Hist. Nat. Rept., vol. 8, p. 352 (type loc.: Egypt). Varanus griseus BOULENGER, 1885, Cat. Lizards Brit. Mus., vol. 2, p. 306. ANDERSON, 1898, Zool. Egypt, vol. 1, p. 134 col. pl. Smirn, 1935, Fauna Brit. India, vol. 2, p. 400. Varanus (Psammosaurus) griseus MERTENS, 1942, Abh. Senck. Mus., Abh. 466, p. 338, pls. 1, 3, 11, 15, figs. 3, 13, 51-52, 91 respectively. MATERIAL EXAMINED (1): CAS 84641. ReMArRKS: According to Smith (1935, p. 401) the scales on the crown of the head are usually larger than those in the neck. In our specimen the lateral nuchal seales are noticeably larger than those on the crown. The dorsum is light tan and is crossed by eight dark brown cross bars on the body, each approximately four seale rows in width, and at least nine eross bars on the tail; there are indications of a number of yellowish cross bars between the dark brown markings. The head is medium to light brown with two parallel dark brown stripes, one originating at the corner of the 454 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH SER. eye, the other below and somewhat behind the eye, both then coursing posteriorly onto the neck where they join with their fellows from the oppo- site side at the middorsal line just in front of the shoulder. The specimen measures approximately 350 mm. in standard length; the tail has been damaged. This species, found widely distributed throughout the Middle East, southwestern Asia, and India, has been recorded from the Shorawuk Plain, at Robat I, on the Afghanistan—Baluchistan frontier (Aleock and Finn, 1896, p. 558; Annandale, 1906, p. 197 [citing previous record]) and in the valley of the Hari-rud (Boulenger, 1889, p. 99). Order SERPENTES FAMILY BOoIDAE Eryx jaculus miliaris (Pallas). Anguis miliaris PALLAS, 1773, Reise Rus., vol. 2, p. 718 (type loc.: Caspian Sea). Eryx jaculus miliaris Borrrcer, 1888, Zool. Jahrb., vol. 3, p. 939. SruLy, 1935, Proc. Boston Soc. Nat. Hist., vol. 40, p. 407. MATERIAL EXAMINED (1): CAS 84638. ReMARKS: This species has been recorded from Afghanistan only once (Boulenger, 1889, p. 101). According to Dr. J. E. T. Aitchison (in Boulen- ger, 1889, p. 101) it was found to be common throughout the Badghis region. Two other species of Eryx have been recorded from this country; EH. ele- gans, described from Afghanistan, and EL. tataricus, recently reported from Afghanistan by Terentjev and Chernov (1949, p. 230). Eryx johni, known from northwestern India, Pakistan (in Baluchistan), and eastern Iran, has not yet been taken in Afghanistan, although it seems likely that it will be found there. Measurements (in mm.) and counts for our female specimen are as fol- lows: scales in 49 longitudinal rows, ventrals 191, subeaudals 19, standard length 396, tail length 34. Family CoLUBRIDAE Coluber karelini Brandt. Coluber (Tyria) karelini BRANDT, 1838, Bull. Acad. Sci. St. Petersbourg, vol. 3, p. 243 (type loc.: southwestern Asia). Coluber karelini Smitn, 1943, Fauna Brit. India, vol. 3, p. 169. Vor. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 455 MATERIAL EXAMINED (7): CAS 84630-84636. Remarks: The identification of these specimens is open to question. The specimens at hand may be divided into two groups. The three smaller individuals, characterized by the color pattern in which there are a series of 40 or more discrete rectangular black cross bars on the dorsum and a series of lateral blotches which alternate with the cross bars, range in their ventral count from 202 to 211, and their subeaudal count from 100 to 102. The four remaining specimens, all larger, have indistinet black cross bars on the body. Among these the ventrals range from 221 to 228, and the subeaudals range from 110 to 118 (for both sexes). In six of the seven specimens, the sixth upper labial is separated from the ocular border by a small subpostocular; in one specimen, however, these two shields are fused and the sixth upper labial borders the orbit. Based upon the structure of the sixth upper labial, which in six of the seven specimens does not border the orbit, it is concluded that these speci- mens must be assigned to C. karelini, although the ventral counts of the four largest specimens fall outside the range for this species. The ventral counts of these large specimens, however, agree with those of C. rhodorachis, a species distinguished from C. karelini by the character of the sixth upper labial, which borders the orbit. Excluding from consideration the character of the sixth upper labial, the three smaller specimens could easily fit the description of C. ventro- maculatus. It must be pointed out that the three species already mentioned, C. kar- elint, C. rhodorachis, and C. ventromaculatus, are known from southwest- ern Asia where their distributions overlap most extensively. They exhibit the same ranges of morphological variation, the same color pattern varia- tions, and are found in similar environmental situations. It seems rather striking, at least to this author, that three species which show so close a morphological and ecological identity, which are sympatric throughout most of their ranges, and which are distinguished, at most, by two variable characters could exist sympatrically and yet be repro- ductively isolated. In regard to the variable morphological characters, I have already noted the possible variation in the subpostocular and sixth upper labial which is exhibited among the specimens at hand. When present, the sub- postocular prevents the sixth upper labial from bordering the orbit and from contacting the subanterior temporal. However, when the subpost- ocular is absent or fused to the sixth upper labial, then this latter shield and the fifth upper labial both border the orbit, and the sixth shield borders the subanterior temporal. Close examination of the ventral counts given by Smith (1943, pp. 167— 456 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. 170) suggests something is amiss which cannot be explained by separating “ventral count groups” into distinet species. According to Smith, C. ventro- maculatus may be characterized by a range of ventral counts from 199 to 211, C. rhodorachis by a range for males of 205 to 229 (252), and for fe- males 218 to 244, and C. karelini by a range of 193 to 212. It should be noted that Smith includes among his counts for C. rhodorachis a single male with 252 ventrals, more than 20 ventrals above the limit for males, and in fact 8 more than among his sample of female specimens. I believe it is self-evident that the difference between 199 and 211, the known range for male C. ventromaculatus (Boulenger, 1894, pp. 400-401) and 205 to 229, the presumed range for males of C. rhodorachis is considerably less than the difference between 205-229 and 252, the range of ventrals for males within the single series. Inso far as C. ventromaculatus and C. karelini are concerned, based upon a comparison of Smith’s data, and the examination of specimens at hand, these species are indistinguishable but for the variable character of the sixth upper labial. It is not my intention here to alter the established nomenclatural or zoological arrangement of these species, but only to point out that their recognition is based upon characters of rather nebulous taxonomic value, and their present partition is not entirely satisfactory. Measurements (in mm.) and counts for our sample of C. karelini are summarized in table IV. TABLE IV Measurements (in mm.) and counts for specimens of Coluber karelini CAS Upper Labials Subpost- Standard Tail number Sex labials in eye ocular Ventrals Subcaudals length length 84630 2 9/9 5, 6 — 222 == 548 = 84631 3} 9/9 5 + 223 118 724 365 84632 2 9/9 5 + 228 1104 670 2334 84633 2 9/9 5 + 221 115 626 224 84634 Q 10/9 5 + 211 100 564 193 84635 3} O59) 5 + 202 == 296 — 84636 3 9/9 5 + 209 1024 185 584+ 2 + = shield present; —- = shield absent. Lytorhynchus ridgewayi Boulenger. Lytorhynchus ridgewayi BOULENGER, 1887, Ann. Mag. Nat. Hist., ser. 5, vol. 20, p. 413 (type loc.: Chinkilok, Afghanistan); 1889, Trans. Linnean Soc. London, ser. 2, vol. 5, p. 102, pl. 11, fig. 1. Smirn, 1943, Fauna Brit. India, vol. 3, p. 190, fig. 58. Vout. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 457 MATERIAL EXAMINED (1): CAS 84639. Remarks: A single female of this deserticolous species was collected. A specimen of Alsophylax tuberculatus was found in its stomach. Measurements (in mm.) and counts of this specimen are as follows: 12 maxillary teeth; 7/8 upper labials; 11 lower labials; nasal divided; 1/1 loreal; 1/1 pre- and 2/2 postoculars; 2/2 suboculars; 2/2 anterior temporals; prefrontal divided; scales in 19 longitudinal rows at midbody; 182 ventrals; anal plate single; 41 dark blotches on dorsum; standard length 334; tail damaged. Psammophis schokari (Forskal). Coluber schokari ForsKAt, 1775, Dese. Anim., p. 14 (type loc.: Yemen). Psammophis schokari BOULENGER, 1896, Cat. Snakes Brit. Mus., vol. 3, p. 157. SMITH, 1943, Fauna Brit. India, vol. 3, p. 363. MATERIAL EXAMINED (1): CAS 84629. Remarks: Only one female of this species was obtained. The color pattern of this specimen is normal except that the dorsomedian pair of black stripes are broken up into series of longitudinally arranged black spots. Measurements (in mm.) and counts of this individual are as follows: seales in 17 longitudinal rows at midbody; 9/9 upper and 10/10 lower labials; nasal divided; 1/1 loreal; 1/1 pre- and 2/2 postoeculars; 2+2+3/ 2424-3 temporals; 184 ventrals; 116 subeaudals; standard length 634; tail length 312. Boiga trigonata melanocephala (Annandale). Dipsadomorphus trigonata var. melanocephalus ANNANDALE, 1904, Jour. Asiatic Soc. Bengal, vol. 73, p. 209, pl. 9, figs. 3—4 (type loc.: Iranian-Baluchistan frontier). Boiga trigonata SmirH, 1943, Fauna Brit. India, p. 349 (in part). MATERIAL EXAMINED (1): CAS 84637. Remarks: Annandale’s variety Dipsadomorphis t. melanocephalus, based upon three specimens with dark heads, was considered by Smith (1948, p. 350) to be a color variety not worthy of subspecific rank. In view of the fact that the dark-headed forms of B. trigonata appear to occupy a geographical area distinct from the more typical color phase of this species, T am inclined to regard it as a distinct subspecies. The specimen at hand has been badly damaged. Its measurements (in mm.) and counts are as follows: 11 maxillary teeth; scales in 21 longitudinal rows at midbody; 8/8 upper labials, the third, fourth and fifth bordering the orbit; 1/1 pre- and 2/2 postoculars; 2+3/2+2 temporals; 230 ventrals; standard leneth (approximate) 470; tail damaged. 458 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. Annandale does not give the exact locality at which his specimens were collected. Smith records B. trigonata from Baluchistan, the North West Frontier Province, and the Transeaspian region, as well as elsewhere in India. Our record appears to be the first for Afghanistan. Family VIPERIDAE Echis carinatus (Schneider). Pseudoboa carinata ScuneEIweER, 1801, Hist. Amphib., vol. 2, p. 285 (based on figure by Russell, 1796, Ind. Serp., vol. 1, pl. 2). Echis carinata Fayrer, 1874, Thanatophidia India, pl. 12. Echis carinatus SmitH, 1943, Fauna Brit. India, vol. 3, p. 487, fig. 154. MATERIAL EXAMINED (2): CAS 84627-84628. REMARKS: The two specimens, both females, agree closely with deserip- tions of this species. Specimen CAS 84628 contained the partially digested remains of an Acanthodactylus c. cantoris in its stomach. Measurements (in mm.) and counts of these specimens are given in table V. TABLE V Measurements (in mm.) and counts for specimens of Echis carinatus CAS Scale Upper Circum- Inter- Standard Tail number rows labials oculars oculars Ventrals Subcaudals length length 84627 36 iPAyAlat abel 8) 181 Bul 415 45 84628 31 13/11 18 10 182 35 477 56 DISCUSSION Afghanistan is a land-locked country bounded by Russian Turkestan on the north, Iran on the east, and the Pakistan provinces of Baluchistan, Sind, and the North West Frontier Province on the south and west. The low lands along the southern borders of Afghanistan lie about 1000 feet above sea level. The level of the land rises toward the north, and espe- cially toward the northeast where altitudes in excess of 25,000 feet are reached. Rainfall throughout the country is meager, especially so in the southern deserts. Extremes of temperature are to be expected, and within one year may range as much as 140 degrees in one locality. The herpetofauna of Afghanistan is largely deserticolous, the genera Vou. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 459 and species being for the most part of Palearctic origin. Genera, such as Agama, Eremias, Phrynocephalus, Uromastix, Varanus, Stenodactylus, Al- sophylax, Echis, Leptotyphlops, Lytorhynchus, Psammophis, and Eryx are all prominently represented in the arid and semiarid regions of south- western Asia, the Transeaspian region, and farther west in the Arabian Peninsula and northern Africa. Other genera found in Afghanistan, notably Teratoscincus and Crosso- bamon, are known from the Transeaspian region and eastward into Tur- kestan. They have entered Afghanistan and eastern Iran, but do not appear to have extended farther into the deserts of southwestern Asia. Three European species of reptiles have entered Afghanistan: T'yphlops vermicularis, Natrix tessellatus, and Vipera lebetina. These species are also known from Transeaspia, Iran, and Iraq. Only a single reptile is endemic to Afghanistan, Eryx elegans. Among the amphibians, two species, Bufa viridis and Rana ridibunda, both of which are widely distributed throughout the Palearctic region, have been found in Afghanistan. In so far as is known, these species are found in the mountains of central and northern Afghanistan where there are sufficient water resources, especially from the melt of the winter snows, to provide for their activities. Recently Smith (1940, p. 382) described a new salamander from the mountain streams above Paghman, at an altitude in excess of 9,000 feet. This salamander, Batrachyperus mustersi, presently known only from the type locality, is related to B. pinchoni and B. tibetanus, both known from the highlands of southwestern China. The fourth amphibian recorded from Afghanistan, Rana sternosignata, has been taken at Arbarp, in the highlands to the immediate west of Kabul, at an altitude of 7,000 feet. The species has been recorded from Sind, Baluchistan, and Kashmir where it has always been found in an aquatic habitat. The relationships of this frog are not certain; Boulenger (1920, pp. 70-71), however, places it next to Rana tibetiana, a species known from the high plateau region of southwestern China. At least two Oriental genera and species have entered Afghanistan from the east, Ptyas mucosus and Boiga trigonata. It would seem remark- able that this latter species, which is adapted to arboreal habitats, should live where such habitats are unavailable. However, other specimens of Boiga trigonata have been found in widely scattered places throughout the Afghan- istan-Baluchistan desert region where there is little or no vegetation, so it is reasonably certain that the snake is not just a casual entrant into this area. In summary, it is evident that Afghanistan has been populated by ele- ments of the Palearctic herpetofauna, including many forms from the Iranian and Turkestan regions, and one or two amphibians from Tibet. 460 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. At least two Oriental genera are also represented. The high-altitude Hi- malayan lizard fauna of northern India and Pakistan does not appear to have made inroads into northern Afghanistan. However, very little col- lecting has been done in that region, and it is by no means certain that future effort will not demonstrate closer affinities between the mountain faunas of northern Afghanistan and India than has been found among the Oriental Indian and Afghan faunal elements to the south. CHECKLIST In the following list of reptiles and amphibians I have mentioned only those forms which have been collected within the political boundaries of Afghanistan or at localities in the vicinity of the international borders between Afghanistan and its neighboring countries. Evaluation of litera- ture records, upon which this list is largely based, has in some eases been difficult. This is especially true of the records of Aleock and Finn, Blan- ford, and Boulenger, which were based upon the collections made in the course of the Persian-Afghanistan-Baluchistan Boundary Delimitation Commission expeditions. Some specimens collected by these expeditions were obtained at border localities; some were obtained at varying distances south of the international border, in Baluchistan, or east of the border, in Iran. I have included in this checklist all records of specimens which were collected at border localities, as for example Koh-i-Malik-Shar, but not from localities several miles distant from the border, as for example Darband. Afghanistan records have been obtained from several sources, most important of which are Alcock and Finn (1896), Boulenger (1889), Smith (1935; 1940; 19438), and Terentjev and Chernov (1949). Class AMPHIBIA Order CAUDATA Family HyNopriipArE Batrachyperus mustersi Smith Order SALIENTIA Family BUFONIDAE Bufo viridis Laurenti Family RANIDAE Rana ridibunda Pallas Rana sternosignata Murray Vout. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 461 Class REPTILIA Order SAURIA Family GEKKONIDAE Agamura persica (Duméril ) Alsophylazx tuberculatus (Blantford) Crossobamon eversmanni (Wiegmann) Eublepharus macularius (Blyth) Gymnodactylus caspius EHichwald Gymnodactylus fedtschenkoi Strauch Gymnodactylus scaber (Heyden) Teratoscincus bedriagai Nikolsky Teratoscincus scincus (Schlegel) Family AGAMIDAE Agama agilis Olivier Agama caucasica (Hichwald) Agama nupta de Filippi Agama ruderata Olivier Agama sanguinolenta (Pallas) Agama tuberculata Gray Phrynocephalus luteo-guttatus Boulenger Phrynocephalus maculatus Anderson Phrynocephalus mystaceus (Pallas) Phrynocephalus ornatus Boulenger Phrynocephalus scutellatus (Olivier ) Phrynocephalus interscapularis Lichtenstein Uromastix asmussi (Strauch) Family ScrNncIpArE Ablepharus grayanus (Stoliczka) Ablepharus pannonicus Fitzinger EHumeces schneideri (Daudin) Eumeces taeniolatus (Blyth) Ophiomorus brevipes (Blanford ) Ophiomorus tridactylus (Blyth) Family ANGUINIDAE Ophiosaurus apodus Pallas Family LACERTIDAE Acanthodactylus cantoris Giinther Eremias acutirostris (Boulenger) Eremias aporosceles (Alcock & Finn) Emerias fasciata Blanford Eremias grammica (Lichtenstein ) EHremias guttulata watsonana Stoliczka Hremias lineolata (Nikolsky) Hremias regeli Bedriaga Eremias scripta (Strauch) Eremias velox persica Blanford Family VARANIDAE Varanus griseus (Daudin) Order SERPENTES Family TyPHLOPIDAE Typhlops vermicularis Merrem 462 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 471 SER. Family LEPTOTYPHLOPIDAE Leptotyphlops blanfordi (Boulenger) Family Boar Eryz elegans Gray Eryz tataricus (Lichtenstein) Eryx jaculus miliaris (Pallas) Family CoLuBRIDAE Boiga trigonata melanocephala Lytorhynchus maynardi Alcock & Finn (Annandale) Lytorhynchus ridgewayi Boulenger Coluber karelini Brandt Natriz tessellata Laurenti Coluber raverigieri Ménétriés Psammophis leithi Giinther Coluber rhodorachis (Jan) Psammophis lineolatus (Brandt) Coluber ventromaculatus Psammophis schokari (Forskal) Gray & Hardwick Ptyas mucosus (Linnaeus) Sphaerosophis diadema (Schlegel) Family ELAPIDAE Naja naja oxiana (Hichwald) Family VIPERIDAE Echis carinata (Schneider ) Vipera lebetina (Linnaeus) Pseudocrastes persica (Duméril, Bibron, Duméril) Order CHELONIA Family TESTUDINIDAE Testudo horsfieldi Gray BIBLIOGRAPHY Aucock, A., and F. FINN 1897. An account of the Reptilia collected by Dr. F. P. Maynard, Captain A. H. McMahon, C.1.E., and the members of the Afghan—Baluch Boundary Commission of 1896. Journal of the Asiatic Society of Bengal, 65:550- 566, pls. 11-15. ANNANDALE, NELSON 1906. Notes on the fauna of a desert tract in southern India. Part I. Batrachians and reptiles, with remarks on the reptiles of the desert region of the North-West Frontier. Memoirs of the Asiatic Society of Bengal, 1:183- 202, pls. 9-10. BEDRIAGA, J. VON 1905. Verzeichnis der von der Central-Asiatischen Expedition unter Stabs- Kapitan W. Roborowski in den Jahren 1893-1895 gesammelten Reptilien. Annuaire du Musée Zoologique de VAcademie Imperiale des Sciences de St. Petersbourg, 10:159-200. Vor. XXIX] LEVITON: SOME REPTILES FROM AFGHANISTAN 463 BoOULENGER, GEORGE ALBERT 1889. Reptiles and batrachians. In: Aitchison, J. E. T., The zoology of the Afghan Delimitation Commission. Transactions of the Linnean Society of Lon- don, ser. 2, vol. 5, Zoology, pp. 94-106, pls. 8-11. BLANForRD, W. T. 1876. Eastern Persia, an account of the journeys of the Persian Boundary Com- mission, 1870-71-72. Vol. II. The zoology and geology. London, 516 pp., 28 pls. Forcart, LOTHAR 1950. Amphibien und reptilien von iran. Verhandlungen der naturforschenden Gesellschaft, Basel, 61:141—156. LANTZ, L. A. 1918. Reptiles from the River Tajan (Transcaspia). Proceedings of the Zoo- logical Society of London, 1918, pp. 11-17, pl. 1. Murray, J. A. 1892. The zoology of Beloochistan and southern Afghanistan. London, 83 pp. SmitH, MALCOLM ARTHUR 1935. The fauna of British India, including Ceylon and Burma. Reptilia and Amphibia. Vol. II. Sauria. London, xiii + 440 pp., 1 pl. 1940. Contributions to the herpetology of Afghanistan. Annals and Magazine of Natural History, ser. 11, vol. 5:382-384. 1943. The fauna of British India, Ceylon, and Burma, including the whole of the Indo-Chinese subregion. Reptilia and Amphibia. Vol. III. Serpentes. London, xii + 583 pp. TERENTJEV, PAUL V., and S. A. CHERNOV 1949. S. A. Operedlitel Presmykaiushchikhsia i Zemnovodnykh. Moscow, 3 ed., 340 pp. (In Russian.) ZUGMAYER, ERICH 1909. Beitrage zur Herpetologie von Zentral-Asien. Zoologischen Jahrbiichern, Abt. Systematik, 27:481—-508. t- PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Fourth Series Vol. XXIX, No. 13, pp. 465-473, 3 figs. May 29, 1959 FROM PIPEFISH TO SEAHORSE — A STUDY OF PHYLOGENETIC RELATIONSHIPS By EARL S. HERALD Steinhart Aquarium, California Academy of Sciences Valuable information about the relationships within the pipefish and seahorse family, Syngnathidae, ean be obtained by careful study of the brood pouch. As is well known, it is the syngnathid male, and not the female, that carries the eggs during incubation. Depending upon the sub- family and genus, the eggs may be attached to the underside of either the abdomen or the tail. They may be attached loosely, imbedded singly in spongy sockets, or crammed into a covered pouch. Bony pouch-pro- tecting plates and pouch folds may either be present or absent. Most ichthyologists are in agreement on Duncker’s (1915) subfamily classifi- cation based upon these criteria (table 1). From this table it will be noted that the abdominal-pouch syngnathids (Gastrophori) do not have a sealed pouch equivalent to the tail-pouch Hippocampus (Urophori). The term Protourophori is given to the hypothetical group of tail-pouch pipefishes corresponding to the Nerophinae. Although not one of this type is known from the fossil record, the other three subfamilies of the Urophori (Solegnathinae, Syngnathinae, and Hippocampinae) must surely have evolved through this stage at some time in their development. Detailed studies of the manner in which the brood-pouch folds close over the eggs have not previously been made. The method of closure (herein designated by the letters BPC) shows to best advantage on males that have had a full clutch of eggs in the pouch for only a few days. [ 465 ] [ Proc. 4TH SER. CALIFORNIA ACADEMY OF SCIENCES 466 AVNIdWVOOddIH ‘fF AVNIHLVNONAS ‘S AVNIH(WNVHYANOG “¢ AVNIHLVYNDAIOG °Z AVNIGIOHLVNONAS °% (9eULIOYdOIN-0jJ0O1d) "TL AVNIHdOWEN ‘TL ({le} Jepun) (usuIOpqge Japun) IHYOHdOU Lf) INOHdOULSVY) UOI}ISOd Yonog pooig pue 33q ‘(sndupno0ddiy) Juesqe 10 (DiNUOLJUADY) JUSSeId saye[d Bur.00101d yonod ‘suruedo [eue-jsod o[surs Surry yonod pooiq po[ees UL pouleju0D sssqq ‘yuesoid Y0q 10 SSULIZAOD Yonod pooiq snouvaiqurIeul 10 soze[d Sul}00}0I1d [eieye, !xtayeu ASuOdsS UL SUIppequI Ysnoiy}, pezelosr sssq ‘SouUBRIqUIOUL SUII9AO) 10 Soye[d BUT}00}J0I1d OU ‘eTBUT Jo doeJINS [V1JU9A UO ssetl AsUOdS UI A910 Yovo WOOL] po}e[OSt ‘A[SUIS pappoquIl s3sqy (Pp) (9) *1a]JO ue 9UO WOdJ poyelOSI Jou nq ‘aTeUI JO dDeJAINS [VIJUBA OF poatoe}}e ATISOO] s3sqq (appiynUbuhig) sassoyvas pun saysyadid ay. fo uUorpayissn1p T qIav EL “” SEAHORSE 467 HERALD: FROM PIPEFISH TO VoL. XXIX] Pied ‘yonod pooiq 34} Jo JUeWIdojeAep [VoTeyOdAH “T “sly mola siellel ahve TSaQNVN IVNIHdOYSN JVNIHdOYN-OLOYd IVNIGIOHLYNONAS 4 >SAYNIHLYN9310S IVNIHdWVHYANO SVNIHLVNONAS SAYLYIIOY{AIOD ©069 sny{oUubuAsos0¢d | |\SNy{OUBUASOIIIY bans Hans ry SS) SsnfouojdaT SVNIdDWVOOdd 1H Fees SNaWwoI0AY{Y49/ | DINUOL{UII VE & SNYLOUBOIDIG eerere) snydwoysAyoo1 J snjyjoydopnasd | faraeesan sndwor0daifye—_ SUMAN YIU ease) Gaia's Ga esleys INIiddV IY SAO IWAS GALYSANI SMe Od oe iF (Cra sao 468 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. At times it may be necessary to examine a series of males of a single species before suitable individuals illustrating the BPC can be found. If the pouch is empty, or if the eggs have reached the advanced embryo stage so that the covering pouch folds have become expanded by the erowing progeny, the character of the BPC may not be fully evident. Given sufficient material, all syngnathine pipefishes with brood-pouch folds ean be assigned to one of the four groups diagrammed in fig. 1. The most generalized type of pouch within the subfamily Syngna- thinae is that of the semi-brood-pouch closure, which is well represented by Corythoichthys. In this genus the eggs are one layer in depth and four to six in width; they are not completely covered by the brood-pouch folds so that the eggs in the center are exposed. Males of the follow- ing species have been examined and all fit into semi-BPC: the Atlantic Corythoichthys albirostris and C. brachycephalus, and the Paeifie C. in- testinalis, C. ocellatus, C. schultzi, C. flavofasciatus, and C. nigripectus. The Atlantic species have pouch-protecting plates and usually have the eges covered to a greater extent than do the Pacifie species which lack the protecting plates. Bryxz, a subgenus of Syngnathus, is similar to Corythoichthys in that it also is a member of the semi-BPC group, but the eggs are arranged in only two rows. From this semi-BPC any of the other types may have developed. If a species of Corythoichthys were to elongate the brood-pouch folds with- out developing thickenings at the free ends, the folds would then be eapable of overlapping as found in overlapping-BPC. This closure method is characteristic of a group of small pipefishes including the At- lantic American Syngnathus dunckeri and the Pacific American Syng- nathus arctus and Syngnathus coccineus. These may now be grouped under a new subgeneric name, Microsyngnathus, herein designated.! If the underlapping edge of the brood-pouch folds were to be turned outwardly back upon itself, this would be the type designated as everted brood-pouch closure. Such closure is characteristic of those pipefishes which are in the direct evolutionary line leading from the primitive open-pouch pipefish to the complex closed-pouch seahorse. This trend, indicated in fig. 1, is shown in detail in fig. 2. The everted-BPC series begins with pipefishes such as the Atlantic Syngnathus elucens and the China Sea Syngnathus argyrostictus. These ean be assigned to the subgenus Parasyngnathus.2 Leptonotus as shown 1Microsyngnathus, new subgenus of Syngnathus; genotype Syngnathus dunckeri Metzelaar; diagnosis: small pipefishes usually less than 100 mm. of the overlapping-brood-pouch closure type; eggs usually maaan two rows. ttt *Parasyngnathus Duncker, 1915, a subgenus of Syngnathus; genotype Syngnathus argyrostictus Kaup; diag- nosis: pipefishes of the genus Syngnathus with everted-brood-pouch closure. Vou. XXTX] HERALD: FROM PIPEFISH TO SEAHORSE 469 , Pouch Hippocampus Caudal |permanent{ Head at absent, ly nearly QQO0 Tail closed right OOO) prehensile exceptfor| angle post-anal | to body pore i Pouch Amphelikturus Gada nae dendriticus Dorsal | vestigial; closed ; fin base 7 Tgi| over slightly Tyeieal raised prehensile posterior] inclined seahorse two rings lateral ridge Trachyrhamphus system serratus LTT SS Micrognathus vittatus eo brevirostris XN Caudal | Pouch Bombonia Intermedi- normal; | open Head spicifer need Tail for in line djarong ridge RON s) complete] with body system Dorsal prehensile length fin base t P not Leptonotus an blainvilleanus TLL mae fe = AAT TLL Typical en NN Syngnathu. lateral Syngnathus Che system e/ucens argyrostictus : SSH Fig. 2. Hypothetical evolutionary sequence from a typical Syngnathine pipe- fish with everted-brood-pouch closure to a typical seahorse. 470 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. by L. blainvilleanus, is perhaps not on the direct line of evolution, for in this genus the lateral trunk and tail ridges are continuous rather than subeontinuous (fig. 3). However, two species previously placed under Syngnathus, viz., S. spicifer and S. djarong, are certainly transitional and should be considered intermediate between Syngnathus and Micro- gnathus. This is indicated by the lateral trunk ridge being deflected ventrad at the anus but not continuous with the inferior tail ridge as is the case in the next genus, Micrognathus. Because of this transitional ridge system, Syngnathus spicifer and S. djarong can be given generic recognition under Bombonia Herre, 1927.* Micrognathus is the next genus in the everted-BPC series. Of the nine species in the subgenus Micrognathus, only two, the Atlantic M. vitatus and the Pacifie MW. brevirostris, have had suitable males that indi- cate that they may be assigned to the everted-BPC series. The other subgenus of Micrognathus is Anarchopterus Hubbs, 1935. It is prin- cipally confined to the Caribbean with only two species, both of which have smooth bodies and lack the anal fin. One species, I. (Anarchopterus) crinigerus, is definitely of the everted-BPC type, and for the other more material is needed. It should be noted that there is a tendency within the species of Micrognathus as well as in Ichthyocampus for the posterior end of the egg-filled pouch to assume an overlapping-BPC although the anterior end will usually be of the everted-BPC type. Trachyrhamphus, as illustrated by the species 7’. serratus, is not ereatly different from JMicrognathus except for two features, the pres- ence of a raised dorsal-fin base, and the loss of the pouch-protecting plates. Amphelikturus, with a single species, A. dendriticus, is the genus of the Syngnathinae most closely related to the typical seahorse. The dor- sal base is raised, the caudal is vestigial, the tail is prehensile, the pouch is closed for the two most posterior rings, and the head is moderately bent. From here it is only a slight jump to the true seahorse, Hippo- campus, which also has the raised dorsal base and the prehensile tail but has lost the caudal fin. The pouch is permanently sealed for most of its leneth with the exception of a small post-anal pore. The head is conspic- uously bent at right angles to the body. There are three genera which have the Micrognathus-Hippocampus type ridge pattern and probably belong to the everted-BPC series. Un- fortunately, suitable males of these three, Yozia, Halicampus, and Hali- ichthys, have not as yet been available. All three have the raised dorsal Becker, disgaoris: pigehches Of the cya BEC tebe ghich fave a, lateral tidve caystear areata ae between Syngnathus and Micrognathus, i.e., lateral trunk ridge deflected ventrad at the anus but not continuous with the inferior tail ridge. Duncker (1915) lists Syngnathus cyanospilus Bleeker as belonging to this deflected ridge group. The single male I have examined appears to have everted-BPC, but definite verification will have to await examination of more specimens. VoL. XXIX] HERALD: FROM PIPEFISH TO SEAHORSE 471 base. Haliichthys has also lost the caudal fin and has gained a prehensile tail. The seahorse subfamily, Hippocampinae, is made up of two genera: (1) Hippocampus, the typical seahorse, which was developed through the everted-BPC series, and (2) the pseudo-seahorse, Acentronura, which obviously did not evolve in the same manner. Both of these fused-pouch genera are alike in having a prehensile tail without a caudal fin, in hav- ing a raised base for the dorsal fin, and in having the head at an angle to the body. Acentronura and Hippocampus both have a ridge system in which the lateral trunk ridge is continuous with the inferior tail ridge. A significant difference, however, is that the lateral tail ridge is entirely missing on Acentronura. Likewise, Acentronura has lateral protecting plates in the brood pouch, these being absent in Hippocam- pus.* In the subfamily Syngnathinae there are only three genera, Uro- campus, Penetopteryx, and Ichthyocampus which have the same kind of ridge pattern as Acentronura. Of these three, Urocampus lacks the pouch- protecting plates, Penetopteryx lacks the dorsal and pectoral fins, but Ichthyocampus has all three structures missing from the others. Pos- sibly then, Ichthyocampus may be the modern representative of a form through which the pseudo-seahorse, Acentronura, has evolved. Of the 18 species in the genus Ichthyocampus only four have a ridge pattern like that of Acentronura: Ichthyocampus bikiniensis from the Marshall Islands, 7. bannwarthi from Suez, I. pawner from the Bahamas, and I. filum from southern Australia and New Zealand. The last species is definitely in the everted-BPC series, but for the others sufficient mate- rial has not been available. In addition to the Acentronura type ridge system, Ichthyocampus has two additional ridge patterns; subgeneric designations are available for all three.® Jchthyocampus pictus is the only other species in the genus that without question can be assigned a BPC group—in this case also the everted type. There are two possi- bilities for the development of the Ichthyocampus-Acentronura group— either they were derived from the main Hippocampus line somewhere between Bombonia and Micrognathus or they arose even earlier than that, before the development of the Parasyngnathus group. Most of the species in the genus Syngnathus belong to the type sub- genus, Syngnathus, characterized by inverted-BPC (genotype Syng- ‘Most of the genera and species in the everted-BPC series have the protecting plates. The exceptions are Yozia, Trachyramphus, and Corythoichthys. 5The three Ichthyocampus subgenera are (1) Larvicampus Whitley 1948, genotype Festucalex runa Whitley = Ichthyocampus filum Giinther, lateral trunk ridge continuous with inferior tail ridge, everted-BPC; (2) Ichthyocampus Kaup 1856, genotype Syngnathus carce Hamilton, lateral trunk ridge deflected ventrad at anus but not continuous with inferior tail ridge, BPC probably everted type; (3) Festucalex Whitley 1931, genotype Syngnathus cinctus Ramsey, lateral trunk ridge ending free near or beyond anal ring, and without ventral deflection, BPC unknown. Another subgenus of Jchthyocampus has been described, viz., Bulbonaricus Pevald,, ale Although it has the Festucalex ridge pattern, its future relationship based on BPC remains o be determined. 472 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. nathus acus l.). The most primitive development of the inverted-BPC series is to be found in such species as Syngnathus californiensis, S. auliscus, S. schlegeli, and S. rostellatus, and in the genus Stigmatopora. In all of these the contacting edges of the brood-pouch folds are only moderately thickened and usually extend slightly, if at all, into the brood pouch. One of the variations of this closure is that found in S. typhle in which the contacting pouch folds are covered with vertical or oblique grooves. All males of this species do not show the character to equal extent, but in some it is so remarkably developed that there may appear to be innumerable small fimbriae attached to the edges of the brood-pouch folds and directed into the pouch. EERC A AP -FirAk Fig. 3. Diagram of the middle part of the body of a hypothetical syngnathid illustrating the terms used in describing this family: (A) anus; (AF) anal fin; (D) discontinuous part of superior tail ridge; (FTAR) first tail ring; (ITAR) inferior tail ridge; (ITR) inferior trunk ridge; (MVR) midventral abdominal ridge; (LTAR) lateral tail ridge; (LTR) lateral trunk ridge; (LTRC) lateral trunk ridge continuous with inferior tail ridge; (LTRR) last trunk ring; (STAR) superior tail ridge; (STR) superior trunk ridge. The majority of the species of the inverted-BPC group belong to the Syngnathus fuscus—S. floridae group, which may be defined as dif- fering from the previous group only in that the folds extend about half way into the center of the pouch, but do not completely divide it. At times it may be difficult to assign a species to one or the other of these groups since the difference between them is only one of degree. Generally speaking, however, the former group often has the eggs crammed into the center area of the pouch as well as on the sides, whereas in the latter croup the eggs are usually on both sides with a partial septum in between, VoL. XXIX] HERALD: FROM PIPEFISH TO SEAHORSE 473 formed by the contacting brood-pouch folds. The septum does not go entirely through the pouch so that there may be eggs in the center portion in the most dorsal section. Other members of this second group are Syngnathus scovelli, 8. pelagicus, S. louisianae, S. springeri, and S. carinatus. One interesting development is found in the Chesapeake Bay population of S. floridae which has the contacting edges of the brood pouch folds covered with small papillae. These are usually visible only under a microscope. The greatest development of pouch division in the inverted-BPC series is to be found in the genus Pseudophallus with three known species limited to Central America. These are P. mindi of the Atlantic and P. starksi and P. elcapitanensis of the Pacific. In this genus the contact- ing surfaces of the brood-pouch folds completely divide the pouch so that the eggs on the two sides of the pouch are entirely separated. From figure 1 it will be noted that the genus Syngnathus breaks down into four subgenera, Syngnathus, Parasyngnathus, Bryx, and Microsyng- nathus, these designations being dependent upon the method of closure of the brood pouch. Bryx and Microsyngnathus represent the end de- velopment within the genus of semi- and overlapping-BPC, respectively. With its inverted-BPC the subgenus Syngnathus has the greatest number of species, but the ultimate development of this type of closure is in an- other genus, Pseudophallus. Parasyngnathus, representing the everted- BPC group, has given rise to a number of genera ending with the true seahorse, Hippocampus. In some manner not yet fully understood, Parasyngnathus and the members of the everted-BPC group are related to the development of the IJchthyocampus - Acentronura evolutionary sequence. LITERATURE CITED DUNCKER, GEORG 1915. Revision der Syngnathidae. Jahrbuch der Hamburgischen Wissenschaft- lichen Anstalten, 32:9-120, 1 pl. 2 TAR. eee yy A | ae % “41 ee 6a cee eric t tyes rae ya aes = = iss . ) ce eee = a a as - - | ae 7 ® ‘ —2 < is . ire i) ong 2 76 a ; ge vite PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES Vol. XXIX, No. 14, pp. 475-508, 10 figs. May 29, 1959 A REVIEW OF THE SNAKES OF THE GENUS PSEUDORABDION WITH REMARKS ON THE STATUS OF THE GENERA AGROPHIS AND TYPHLOGEOPHIS (SERPENTES: COLUBRIDAE) BY ALAN E. LEVITON Department of Herpetology California Academy of Sciences AND WALTER C. BROWN Natural History Museum Stanford University, California Recent field work by one of the authors’ on Negros Island, Philippine Islands, resulted in the collection of a series of specimens of Pseudorabdion oxycephalum Giinther and Pseudorabdion mcnamarae Taylor. Close ex- amination of the material suggests that P. orycephalum is more similar to recognized species of T'yphlogeophis Giinther, and P. mcnamarae to recog- nized species of Agrophis Miiller than is either to Pseudorabdion longiceps [= Rabdion torquatum, type species of the genus Pseudorabdion|. These findings have led us to reevaluate the status of these genera and of their included species. 1 Dr. Brown visited the Philippine Islands on two occasions, during 1954-1955 and again during 1957. Asa result of his activities, and with the generous assistance of the staff of the Department of Biology, Silliman Uni- versity, Dumaguete, Negros Island, a considerable collection of amphibians and reptiles has been amassed. These specimens have been deposited at Stanford University. [ 475 J 476 CALIFORNIA ACADEMY OF SCIENCES [Proc. 4TH Ser. In the course of this study we have become indebted to a number of per- sons for the loan of specimens under their care and for other courtesies. Dr. Lothar Foreart, Basel Museum; Dr. Jean Guibé, Paris Museum; and Dr. H. W. Parker and Mr. James C. Battersby, British Museum; have gen- erously aided us by examining type specimens and answering our numerous questions regarding these materials. Mr. Charles M. Bogert, American Mu- seum of Natural History; Dr. Doris M. Cochran, United States National Museum; Mr. Arthur Loveridge and Dr. Ernest E. Williams, Museum of Comparative Zoology, Harvard College; and Dr. Robert F. Inger, Chicago Natural History Museum, have made available to us their specimens of Pseudorabdion. The investigations of the junior author in the tropical forests of the Philippine Islands, upon which this paper is in part based, have been sup- ported since 1955 by grants from the National Science Foundation. Prior to 1955 the work was aided by a grant from the United States Educational Foundation in the Philippines. The illustrations, with the exception of figure 7, were prepared by Mr. Walter Zawoski of the Stanford Research Institute. Figure 7, that of Pseudorabdion sarasinorum, was prepared by Mr. W. Schlier of the Basel Museum. ABBREVIATIONS The following abbreviations are used throughout this work: CAS = California Academy of Sciences, San Francisco. CNHM = Chicago Natural History Museum, Chicago, Illinois. MCZ = Museum of Comparative Zoology, Harvard College, Cambridge, Massachusetts. SU = Natural History Museum, Stanford University, California. USNM = United States National Museum, Washington, D. C. TERMINOLOGY Several terms used in this paper require explanation or definition: Caudodorsal scale reduction — the reduction of caudodorsal scales from six to four results from the fusion of the second and third scale rows; the position of this reduction is expressed in terms of the subeaudal plate, num- bered by counting posteriorly from the anal aperture, opposite which the reduction takes place (indicated by the number in brackets []): e.g., 6 (2+ 3 [8]) 4. Loreal — any shield immediately behind the nasal shield which is distinet from the prefrontal or labial shields and which, even when it borders the orbit, does not prevent the preocular, if present, from also bordering the orbit. VoL. XXIX] LEVITON AND BROWN: THE GENUS PSEUDORABDION 477 Preocular —a shield immediately anterior to and bordering the orbit; when this shield is present the prefrontal does not enter the orbit. Standard length —the distance from the tip of the snout to the anal aperture. Subcaudals — these paired and enlarged shields on the underside of the tail are counted from the first overlapping pair immediately behind the anal aperture to the last overlapping pair just anterior to the tip of the tail. Supraocular —a shield which borders the orbit from above and which normally prevents the frontal from bordering the orbit; this shield may be very small and occasionally fused to the postocular. Ventrals — Dowling’s method (1951, pp. 97-98) has been used in count- ing the enlarged shields on the underside of the body. Genus PSEUDORABDION Jan Rabdion DumMeértIL, 1853, p. 441 (in part). DUMERIL, Brsron and DumeErIL, 1854, p. 115 (in part). Pseudorabdion JAN, 1862, p. 10 (type species Rabdion torquatum DUMERIL, BIBRON and DumeértIn, 1854, by monotypy). Oxycalamus GUNTHER, 1864, p. 199 (type species Calamaria longiceps CANTOR, 1847, by monotypy). Typhlogeophis GUNTHER, 1879, p. 77 (type species Typhlogeophis brevis GUNTHER, 1879, by monotypy). Pseudorhabdion BouLENGER, 1885, p. 389 (erroneous subsequent spelling). Rhabdion BoEettrcER, 1886, p. 106 (erroneous subsequent spelling). Pseudorhabdium BouLENGeER, 1894, p. 328 (erroneous subsequent spelling). Agrophis MUxuer, 1894, p. 827 (type species Agrophis sarasinorum MULLER, 1894, by monotypy). Typhlogeophus Casto DE Eirera, 1895, p. 425 (erroneous subsequent spelling). DeEFINITION: Head not distinet from neck; internasals present, smaller than enlarged prefrontals; loreal present or absent; preocular generally absent (present in P. longiceps) and prefrontal usually borders orbit; an- terior temporals absent; 5 or 6 upper labials; scales in 15 longitudinal rows throughout, smooth, without apical pits; subeaudals paired; anal plate sin- ole; pupil round; maxillary teeth 8 to 21; hypapophyses absent on posterior vertebrae; hemipenes and suleus spermaticus forked (in those species whose hemipenes have been examined). Remarks: The genus Pseudorabdion was established by Jan (1862, p. 10) to accommodate a single species, Rabdion torquatum Dumeéril, Bibron, and Duméril, which was formerly included with Rabdion forstent Duméril, Bibron, and Duméril in the nominal genus Rabdion Duméril [= Rhabdo- phidium Boulenger; type species R. forstent by subsequent selection by Leviton, 1958, p. 44]. Ginther (1864, p. 199) proposed the genus Oxycalamus for a single 478 CALIFORNIA ACADEMY OF SCIENCES | Proc. 4TH SER. species, Calamaria longiceps Cantor. In 1873, he transferred Rhabdosoma oxycephalum Giinther (1858) to this genus. Boulenger, in 1894, referred Oxycalamus to the synonymy of Pseudorabdion |[=Pseudorhabdium Bou- lenger| having shown that Calamari longiceps Cantor and Rabdion tor- quatum Dumeéril, Bibron, and Duméril were conspecific. In 1917, Taylor described a new species of Psewdorabdion, P. mcnam- arae, and in 1922 followed with a second new species, P. minutum. Thus four recognized species were assigned to Pseudorabdion: P. longiceps, P. mc- namarae, P. minutum, and P. oxrycephalum. Pseudorabdion longiceps, type species of the genus, is a distinctive species characterized by the presence of a small preocular, absence of the loreal, a large eye, 12 maxillary teeth, a rather narrow conical head, and a bifurcated hemipenis. In some of these characters P. longiceps agrees with its congeners; it differs in the details of ornamentation of the distal hemi- penial walls, the presence of a preocular shield, the greater number of max- illary teeth, and in the larger eye. Pseudorabdion mcnamarae has a loreal shield, thereby differing from both P. longiceps and P. oxrycephalum. However, fusion of the loreal and the prefrontal would result in a pattern of head scutellation almost identical with P. orycephalum. The supraocular and postocular shields are fre- quently fused together. The hemipenes of P. mcnamarae are identical in ornamentation to P. oxycephalum; they differ only in absolute length, P. mcnamarae having the shorter organ. Furthermore, both P. mcnamarae and P. oxrycephalum have 7 or 8 maxillary teeth, 4 or 5 less than in P. longiceps. Little need be said in regard to P. minutum at this time. It is very close to P. mcnamarae, differing only in having fewer subeaudals and the absence of a light nuchal collar. Pseudorabdion oxrycephalum appears to be the most highly adapted species in the genus. The loreal and preocular are both absent. The supra- ocular and postocular are fused together and both are fused to the ocular shield. The head is rather broad and the eye is very small (in some speci- mens the eye appears only as a slight darkened bulge beneath a thickened ocular shield). There are 8 maxillary teeth, and the tail is very short. Over-all similarity among the four species of Pseudorabdion would in- dicate they are congeneric. In 1879, Giinther erected the genus T'yphlogeophis for a presumably new species of blind burrowing snake, 7. brevis. Exeept for the character of the eye, which is said to be absent in this species, we believe it is indis- tinguishable from P. oxycephalum. Mr. J. C. Battersby and Dr. H. W. Parker of the British Museum have kindly reexamined the type of 7. brevis and concur in this view. However, they have stated (in litt.) that the con- Vout. XXIX] LEVITON AND BROWN: THE GENUS PSEUDORABDION 479 dition of preservation of the specimen does not permit a definite statement regarding the presence or absence of the eye. In 1922, Taylor described a second species of T'yphlogeophis. In T. ater the eye was said to be clearly visible, although “hidden” under the ocular shield. We have examined the type and find the condition of the eye to be similar to that in P. orycephalum, specimens of which had not been seen by Taylor. Differences in head scutellation and dentition suggest 7’. ater is specifically distinet from P. oxycephalum, but not generically distinct. At the time Miiller (1894) proposed the nominal genus Agrophis, based upon a single species of A. sarasinorum, he suggested it might be con- generic with Oxrycalamus Giinther. Agrophis sarasinorum is a distinctive species possessing 1) a loreal shield, 2) 14 subequal maxillary teeth, 3) a large divided nasal shield, and 4) in having the mental separated from the anterior chin shields. The loreal shield is similar in appearance to that in P. mcnamarae and there are only two more maxillary teeth than are found in P. longiceps. The distinguishing characters of A. sarasinorum are, there- fore, items 3 and 4. The hemipenes of this species are unknown. If they are found to be similar to those known among other species of Pseudorab- dion, aS we suppose, Agrophis should be placed into the synonymy of Pseudorabdion. In the present work we tentatively regard Agrophis as a synonym of Pseudorabdion. Two additional species, usually assigned to Agrophis, Geophis albonu- chalis Giinther, and A. saravacensis Shelford, are more similar to P. mc- namarae than to A. sarasinorum. An elongate loreal is present in both species, the mental is in contact with the anterior chin shields, and the nasal shield, although somewhat larger than in species of Pseudorabdion, is un- divided. There is a marked similarity between these species and Pseudo- rabdion in the relationships of the head shields, although this may be the result of convergence among morphologically adapted groups; G. albonu- chalis agrees rather closely with 7. ater, and A. saravacensis with P. mc- namarae and with a species described in this paper as new, P. taylori. Agrophis saravacensis differs from all mentioned species in the large num- ber of maxillary teeth, 18 to 21 according to Mr. J. C. Battersby (in litt.). The dentition of G. albonuchalis is not known. Based upon the evident agreement in scale patterns we are led to conclude that G. albonuchalis and A. saravacensis should be assigned to Pseudorabdion. Two new species are described in this paper, P. montanum and P. tay- lori. Of these, Pseudorabdion taylori is of special interest inasmuch as it appears to be intermediate between P. mcnamarae and the Bornean species assigned to Agrophis. There is agreement with P. mcnamarae in the denti- tion. However, in the larger size of the nasal shield, P. taylori approaches A. saravacensis. The ornamentation of the distal portions of the hemipenes 480 CALIFORNIA ACADEMY OF SCIENCES [ Proc. 4TH SER. in P. taylor differs from other species of Pseudorabdion. The differences are of details and do not affect the generic assignment of the species. In summary, we conclude: 1) P. longiceps, P. mcnamarae, and P. oxy- cephalum are presumed to be congeneric as previously held (Auct.); 2) Typhlogeophis brevis and P. oxycephalum are probably conspecific and Typhlogeophis must be placed into the synonymy of Pseudorabdion; 3) Typhlogeophis ater must be transferred to Pseudorabdion; 4) Agrophis sarastnorum is probably congenerie with P. mcnamarae; 5) Geophis albo- nuchalis and A. saravacensis are both similar to P. mcnamarae and must be transferred to Pseudorabdion. In the following pages each of the recognized species now included in Pseudorabdion is diseussed in some detail. Two new species are described. Of the twelve nominal species placed in the genus, nine are recognized: Geophis albonuchalis Ginther = Pseudorabdion albonuchalis (Giinther) Typhlogeophis ater Taylor = Pseudorabdion ater (Taylor) Typhlogeophis brevis Ginther = Pseudorabdion orycephalum (Giinther) Calamaria longiceps Cantor = Pseudorabdion longiceps (Cantor) Pseudorhabdium mcnamarae Taylor = Pseudorabdion mcnamarae Taylor Pseudorhabdium nunutum Taylor = Pseudorabdion mcnamarae Taylor Pseudorabdion montanum Leviton and Brown, new species Rhabdomosa oxycephalum Giinther = Pseudorabdion oxycephalum (Giin- ther) Agrophis sarasinorum Muller = Pseudorabdion sarasinorum (Miller) Agrophis saravacensis Shelford = Pseudorabdion saravacensis (Shelford) Pseudorabdion taylori Leviton and Brown, new species Rabdion torquatum Duméril, Bibron, and Duméril = Pseudorabdion lon- giceps (Cantor) KEY TO THE SPECIES OF PSEUDORABDION la. Loreal shield absent; prefrontal in contact with upper labials. 2a. Preocular present; supraocular present; internasal not in contact with upper labials: maxillany