I B RARY OF THE UNIVERSITY OF ILLINOIS FI Return this book on or before the Latest Date stamped below. A charge is made on all overdue books. University of Illinois Library 7K}- ROVio" '952 jot 1949 M32 590.5 f X UBKAKt 1 ZOOLOGICAL SERIES OF FIELD MUSEUM OF NATURAL HISTORY Volume 24 CHICAGO, NOVEMBER 20, 1943 No. 31 A STUDY OF THE TORRENT DUCKS BY BOARDMAN CONOVER RESEARCH ASSOCIATE, DIVISION OF BIROS This paper is based on a study of 144 specimens of torrent ducks, genus Merganetta, 45 from my own collection and 99 (all those available) borrowed from leading American museums. These specimens represent all the known forms. For their kindness in making these loans I am indebted to the American Museum of Natural History, Carnegie Museum, the Museum of Comparative Zoology, the United States National Museum, the Philadelphia Academy of Sciences, and the Museum of Vertebrate Zoology, University of California. With full descriptions and colored plates of all the forms except berlepschi available in Phillips' "A Natural History of the Ducks," it has not been thought necessary to repeat such details in this paper. Instead I have limited myself to pointing out the dis- tinctive characters between forms, and some features and variations not dealt with in that work. The torrent ducks seem to fall into three divisions, a northern and a southern one, each containing a single race, and a central group of four races. While all the varieties are undoubtedly closely related, the northern and southern forms differ much more from the central group and each other, than the four central varieties do inter se. Since there also appears to be a rather extensive gap in distribution between the northern, the central and the southern divisions, it seems that the relationships are more clearly shown by separating the genus into three species than by including all six forms as races of one species. By doing this we have a northern and a southern monotypic species and a central one comprised of four races. No. 538 345 346 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. 24 The resulting arrangement for the genus is shown in the following table: Merganetta colombiana Des Murs Merganetta leucogenis leucogenis Tschudi turneri Sclater and Salvin garleppi Berlepsch berlepschi Hartert Merganetta armata Gould Merganetta colombiana Des Murs Merganetta colombiana Des Murs, Rev. Zool., p. 179, 1845 — no type locality given = Colombia by inference. Range. — The northern Andes from the State of Merida, Vene- zuela, south through Colombia (all three ranges and Santa Marta) to central Ecuador (Chimborazo and Rio Pastaza). From central Ecuador to northern Peru (Molinopampa and Leimabamba, Ama- zonas) there seems to be a stretch of the Andes from which no merganettas have been recorded. Males differ from those of the forms of leucogenis and armata by having the mantles brownish gray (not black, black streaked with white, or white) finely vermiculated with white. Nearest to typical leucogenis but smaller (wings 142-155 as against 165-176 mm.), and the dark shaft stripes to the feathers of the breast are narrower and lighter in color with the olive brown edges1 to the scapulars and tertials, on the average, lighter, less reddish. While the females of armata and of the forms of leucogenis are indistinguishable, those of colombiana can be easily separated. The under parts are much lighter, the color being more ochraceous (Ochraceous Tawny of Ridgway) than deep ferruginous. The pos- terior parts of the cheeks and sides of the neck are also of this ochraceous color whereas in the other species these parts are gray very finely barred with white. The upper parts have a somewhat lighter appearance, the dark centers to the scapulars and tertials being duller, less blackish, and therefore less in contrast to the lighter edges. This latter feature, however, is not very diagnostic. Wings 133-140 mm. The immatures, like the females, differ from those of leucogenis and armata, although those of the two latter are alike. In colombiana the posterior parts of the cheeks and the sides of the neck are white 1 In his description of the adult male Phillips writes: "The scapulars have . . . brown or pale gray edges." All the specimens examined have olive brown not pale gray edges. . /V/S7. A/' £ 1943 TORRENT DUCKS — CONOVER 347 (occasionally with slight specklings of dusky) while in the other species these parts are heavily sprinkled with gray. In the first or immature plumage both sexes are alike. The upper parts, including the wings, are much as in the adult female but duller, ~^and the lower back, rump and upper tail coverts are more broadly barred with dusky and white. On the under parts, however, where the female is orange red, the immature is pure white except for the sides and flanks which are broadly barred with dusky. The first tail feathers are dusky, broadly barred or irregularly streaked and spotted with white and when newly grown still have the downy rectrices attached at the end. The downy rectrices, however, break off shortly after the first immature feathers attain their full growth. These first feathers seem to be shed about the time the adult plumage begins to make its appearance and are replaced by a second set, which is dusky with a narrow band of white vermiculations along the edges of the webs. The third and final set of dusky adult feathers without white markings does not seem to be grown until some time after the fully mature plumage has been attained. In none of the merganettas do young males pass through a female stage of plumage with the ferruginous-colored under parts, as has been suggested. Instead the series of immatures examined shows clearly that they pass from the white-breasted stage directly into the adult plumage. In the young males of this species the first adult feathers to make , their appearance are the dusky, olive-brown-edged ones of the lower mantle, scapulars and tertials about in the order named. Some of these are present even before the primaries are fully grown, but in these early ones the olive brown edges are much lighter than they become later. Next to appear are the black-striped feathers of the breast, together with the dusky and white vermiculated ones of , the flanks, rump and upper tail coverts. At about this time the < black feathers of the head and neck also come in. Specimens of immature female torrent ducks are much rarer than those of males. From the four young females of this species examined (against twelve males) it appears that the females molt very quickly into the ochraceous-breasted plumage of the adult. The downy young are colored as follows: Top of head, back of neck and rest of upper parts including wings, and flanks blackish . brown. Sides of head, including superciliary stripe, sides of neck v and under parts white. A stripe through the eye and a spot about the ear blackish brown. There are white bars across the wings and 348 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. 24 a broad stripe of the same color down each side of the back from the base of the wing to the rump. In the middle of the upper back is a large white spot. The tail is dusky, broadly tipped with white. The breeding season of colombiana appears to be very irregular. Downy young have been taken in October and eggs in November. I have examined fully feathered immatures with the primaries not yet grown, taken in the months of January, March, April and September. Because of this irregularity the sixteen immature specimens examined do not give a clear picture as to the age at which this duck becomes adult, but from my study of the immatures of armata it probably does not take over a year. The size of the wing spurs in both males and females appears to increase with age, as do those on the legs of some Gallinae. The spurs in the females, also, are much smaller than in the males. These two facts are often useful in judging the sex of immatures, and in ascertaining what variations in plumages are due to age and what to race. Specimens examined. — VENEZUELA: Merida, 1; La Chuchilla, 2; Nebados, 2; Rio Alba, 2; Laguna Pan de Azucar, 1. — COLOMBIA: "Bogotd," 6; La Plata, Huila, 5; El Roble, Quindio Andes, 1; Rio Toche, east Quindio Andes, Tolima, 1; Cali, 1; Cauca Valley, 1; Salento, Cauca, 2; Cumbal, Narino, 1; Paramo de Purace, Cauca, 1; Munchique, Cauca, 3; west of Harmonde, Cauca Valley, 1. — ECUA- DOR: Oyacachi River, 3; upper Rio Pita, 4; above Tumbaco, 1; Rio Tumbaco, 2; Mount Chimborazo, 1; Antisana, 1; Cerro Galeras, Rio Pauchsi Yacu, 1; Rio San Pedro, near Tumbaco, 1; Rio Quijos, 1; Rio Ambato, 1; Cumbaya, 1; Sumaco, 1; Papallacta, 1; Tunga- ragua, Rio Ambato, 1 ; Rio Shubol, Chimborazo, 1 ; Cerro Mojanda, Imbabura, 1; Rio Mira, 1; La Palmera, near Bafios, Rio Pastaza, 1; San Francisco, Rio Pastaza, 1; Rio Guaillabamba, Pichincha, 1; Rio Chalpi, 1. Merganetta leucogenis Tschudi In the males all forms of leucogenis can be distinguished from armata by having the cheeks, throat and foreneck white and from colombiana by having the mantle black (sometimes washed with olive) or black streaked with white, not brownish gray finely vermic- ulated with white. Females can always be distinguished from the same sex in colombiana by having the posterior half of the cheeks and the sides of the neck gray finely barred with white instead of the same color 1943 TORRENT DUCKS— CONOVER 349 as the foreneck. However, in this species there is great variation in the color of the under parts, although the lightest of thirteen specimens examined is considerably darker than any of colombiana. In leucogenis the under sides vary from a deep to a pale ferruginous (Kaiser Brown to Cinnamon-Rufous of Ridgway). At first I thought that this paler coloration was probably due to immaturity, but on closer examination I found that some of the lighter-colored specimens had well-developed wing spurs while in one of the darkest females not only were these spurs very small but there were still some of the white immature feathers remaining on the abdomen. Owing to this variation lighter-breasted specimens of leucogenis are readily separable from armata, all the females of which seem to be a deep ferruginous, but darker examples are not. Immatures of this species differ from those of colombiana as pointed out under that species, but cannot be told from those of armata, except that the males can be distinguished as soon as the adult facial markings begin to appear. Merganetta leucogenis leucogenis Tschudi Anas leucogenis Tschudi, Arch. Naturg., 9, (1), p. 390, 1843 — Mana Rima- cunan, sources of the Aynamayo, Junln, Peru.1 Range. — Andes of northern (Molinopampa, Amazonas) and central (Maraynioc, Junin) Peru.2 The males of this race differ from those of turneri and berlepschi in the much lighter breast and abdomen, the ground color of these parts being whiter and the dark shaft streaks narrower and perhaps not so blackish, especially about the chest and upper breast. They differ from those of garleppi and also berlepschi by having the mantle black (sometimes washed with olive), not streaked with white, and by having the edges to the scapulars and tertials reddish buff, not buffy gray or white. However, as is pointed out under that race, an occasional specimen of garleppi has these edges light reddish buff, but the mantle is streaked with white. Wings 165-176 mm. Females are indistinguishable from those of the other races, except that those of the more southern forms seem to average larger. Wings 152-157 mm. Only one immature specimen was available for examination. It differs from those of colombiana only by having the posterior parts 1 Cf. Tschudi, Fauna Peruana, 1845-46, p. 312. 2 The records given by Phillips for this form from Cuzco (Urubamba Valley, Rio Keatin, etc.) belong under turneri. 350 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. 24 of the cheeks and sides of the neck finely speckled with dusky. Although this specimen is so young that the primaries are just beginning to grow, the edges to the scapulars already have a faint olive brownish tinge indicating that it is a male. No specimens of downy young were available. This race undoubtedly intergrades with turneri probably just south of Junin. Specimens examined. — PERU: Molinopampa, Amazonas, 1; Leima- bamba, Amazonas, 1; Huanuco, Huanuco, 1; Panao Mountains, Huanuco, 1; Huanuco Viejo, Huanuco, 1; Junin, 1; Maraynioc, Junin, 3; Chipa, Junin, 4; Obrajillo, Canta, I;1 Oroyo, Rio Mantaro, Junin, 4. Merganetta leucogenis turneri Sclater and Salvin Merganetta turneri Sclater and Salvin, Proc. Zool. Soc. Lond., p. 600, (1869), 1870— Tinta, Peru. Range. — Known at present only from southern Peru in the upper Urubamba Valley south of 12° S. lat., from the sources of the Rio Inambari in the Department of Puno, and from Sumbay2 in the Department of Arequipa. Males differ from those of berlepschi and garleppi by having the mantle black (not streaked with white) and by having the edges to the scapulars and tertials reddish buff, not white or buffy white. They differ from those of typical leucogenis and also garleppi by having the under parts much darker. However, there is great variation in this feature, the type, as shown in Phillips' plate, probably representing the dark extreme. It is said to have the chest, flanks, and under tail coverts black and the lower breast and abdo- men dark reddish gray broadly streaked with black. Nine adult males have been examined from southern Peru. In all these the black flanks are lacking and the nearest approach to black chests is found in three examples from the Rio San Gaban, which have black collars at the base of the white forenecks varying from one- eighth to three-eighths of an inch in width. In the two of these three specimens which have the widest collars the vents and under tail coverts are black, but in none of the other seven examples. 1 This specimen, an adult male, is somewhat darker below than normal, approaching the lightest examples of turneri. 2 Chubb (Ibis, 1919, p. 276) records three specimens from this locality and says that he has compared the only male with the type of turneri and finds it even darker. 1943 TORRENT DUCKS— CONOVER 351 The chests, breasts and abdomens of eight of these nine birds, how- ever, are much more broadly streaked with black than in typical leucogenis or garleppi, and the ground color varies from light grayish white to light dirty buff. In one of two specimens from Limbani (the other is more typical of turneri} the under parts are no more broadly streaked with black than in garleppi or some examples of the typical race. The ground color, however, is a rusty buff more reddish than in any of the others, which may be due to immaturity, as the wing spurs are very small. The mantle is faintly streaked with white, and, as the locality, Limbani, is near the Bolivian border, the peculiarities of this specimen probably indicate an approach to the next race garleppi. Wings 168-175 mm. Females and the only immature available are not distinguishable from those of the other races. Wings of females 155-161 mm. Downy young show no difference from those of colombiana. Specimens examined. — PERU: Tocopoqueu, Ocabamba Valley, 2; Huracondo Canyon, 4; Lucma, Cosireni Pass, Urubamba Valley, 1; Keatin (Quiton) River, 2; Limbani, Puno, 2; Rio San Gaban, Puno, 4; Rio Sandia, Puno, 2. Merganetta leucogenis garleppi Berlepsch Merganetta garleppi Berlepsch, Orn. Monatsb., 2, p. 110, 1894 — Cocotal, Bolivia.1 Range. — The eastern Andes of Bolivia south probably to about 20° S. lat. So far as I have been able to ascertain no merganettas have been recorded from western La Paz or Oruro and from the rather arid character of that part of Bolivia and the lack of perma- nent streams it would seem probable that torrent ducks are not found there. Phillips gives records for this race from Yura, southern Peru, and several localities in the Province of Tucuman, Argentina. From present knowledge there is not much doubt that the Peruvian records refer to turneri and the Argentine ones to berlepschi. Males are very similar to those of typical leucogenis on the under side, but perhaps the ground color of the feathers of the chest, 1 No locality by this name can be found on available maps. Dr. John T. Zimmer, of the American Museum of Natural History, has suggested in a letter to me that Berlepsch may have misread the original label, mistaking a capital L for a C and that the type locality is really Locotal. In support of this he called my attention to the facts that the type specimen was taken on March 14, 1891, and that in January of that year Garlepp, the collector of the type, was known to have collected at San Jacinto in the vicinity of Locotal and later, in July of the same year, was on the Rio San Mateo where Todos Santos is situated, which is in the same general area. 352 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. 24 breast and abdomen is a little more buffy. They differ from the typi- cal race and turneri by having the black feathers of the mantle edged with white, and by having the edges of the scapulars and tertials buffy gray (sometimes almost white) not reddish buff. In this latter feature, however, there seems to be some variation, one out of the five adult males examined having these edges light reddish buff, approaching some specimens of typical leucogenis. The mantle, however, was streaked with white. This form differs from berlepschi by its much lighter under parts (this feature distinguishes it also from turneri) and by the somewhat darker edges of the scapulars and tertials, which are whiter in the more southern race. Wings 166- 175 mm. A female is the same as in the other races. Wing 153 mm. Only two immature specimens are available, both males, well along toward the adult plumage. They do not seem to differ from like stages in the other forms. Wings 175-179 mm. No downy young were examined. Specimens examined. — BOLIVIA: Incachaca, Cochabamba, 5; Yungas de Cochabamba, 1; Omeja, Yungas, 1; western Bolivia, 1. Merganetta leucogenis berlepschi Hartert Merganetta berlepschi Hartert, Nov. Zool., 16, p. 244, 1909 — near Tucuman, 1,800 meters, Argentina (type examined). Range. — From southern Tarija (Rio Narvaes), Bolivia, south in the easternmost Andes of Argentina to Tucuman and northern Catamarca (Lago Blanco). The drakes of this race have the under parts dark as in turneri and seem to show the same great variation. Dabbene (El Hornero, 4, p. 34, pis. 3 and 4, 1927) has called attention to this, but has ascribed the differences to age rather than individual variation. Specimen number 1 (and 1, a) shown in his plate 3 is of course an immature. However, from the variations I have seen in specimens of turneri and armata (which always has a black chest) I believe the differences in coloration of the other eight male specimens shown in Dabbene's plates are due entirely to individual variation, although it may be that sometimes a more reddish ground color to the breast is due to immaturity. The type in the Tring Collection of the American Museum of Natural History has been examined and proves to belong to the dark extreme. An adult male (with large wing spurs) from Lago Blanco, Province of Catamarca, is much lighter, having the under parts 1943 TORRENT DUCKS — CONOVER 353 fairly heavily streaked with black on a rather light buff back- ground. This locality is the most western and almost the most southern recorded for this race, but the specimen shows no tendency toward armata in the dark markings of the head and neck. Indeed, on the under side it approaches garleppi more nearly, in coloration. Another specimen from Tarija, Bolivia, in the author's collection, has the chest and upper breast very heavily streaked with black, only the extreme edges of the feathers being white, but the flanks, abdomen and under tail coverts are much lighter. The feathers of the mantle, scapulars and tertials are edged with white. This race is simply turneri with white edges to the feathers of the mantle and white or buffy white (not reddish buff) edges to the scapulars and tertials in the males. From typical leucogenis it can be told by the above characters and by its darker under parts, which latter feature also distinguishes it from the adjoining race garleppi. Wings 175-184 mm. The single female examined has deep ferruginous under parts. Wing 166 mm. No immatures or downy young were available for examination. While on a collecting trip in Argentina I was shown by the col- lector Rodriguez a photograph of what he said was the nesting site of a pair of torrent ducks in Tucuman. It was about seventy-five feet above a stream on the face of a steep cliff, the nest itself being hidden beneath a bush. Specimens examined. — BOLIVIA: Rio Narvaes, Tarija, 1. — ARGEN- TINA: Tucuman (the type), 1; Rio Norco, Tucuman, 1; Lago Blanco, Catamarca, 2. Merganetta armata Gould Merganetta armata Gould, Proc. Zool. Soc. Lond., 9, p. 95, 1841 — "Andes of Chile, lat. 34°-35°"= Province of Colchagua. Merganetta fraenata Salvadori,1 Cat. Bds. Brit. Mus., 27, p. 458, pi. 5, fig. 1, 1895— "Central Chile." Range. — The Andes of Chile and Argentina2 from Coquimbo (Balala, Rio Turbio) and Mendoza (Valle Hermoso) south to 1 Merganetta fraenata Salvadori is supposed to differ from armata by having the black stripe from the throat to the eye extending to the crown, and by having the under parts paler reddish with the black shaft streaks broader. Examination of twelve adult males shows that all these characters are due to individual rather than geographic variation. 2 Records from Jujuy and Tucuman in northwestern Argentina as given by Phillips and others belong under Merganetta leucogenis berlepschi. 354 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. 24 Chubut (Rio Cholila and Rio Epuyen), Aysen (Rio Nirehuao, 46° S. lat.) and Tierra del Fuego (Lago Fagnano). This form has been given full specific status not only because of the very different throat and facial markings of the male (which seem to be deep-seated characters since they are the first features of the adult plumage to make their appearance in the immature), but also because its range appears to be entirely separated from that of the forms to the north and east. From M. 1. turneri and M. I. garleppi it is cut off by the desert region in northern Chile and western Bolivia, while from M. 1. berlepschi it is separated by the extension of this same arid region into northwestern Argentina in the provinces of Los Andes, Catamarca (western part) and La Rioja. The males of this species can always be distinguished by their black throat and the black stripes from the eyes to the throat and down the white foreneck to the black chest. This stripe on the fore- neck, however, is sometimes partially obsolete. Above, the mantle is much whiter (often practically pure white) than in berlepschi or garleppi and the edges to the scapulars and tertials are pure white and broader than usual in those forms. The chest is always black but in the twelve specimens examined there is considerable variation in the coloration and markings of the breast, flanks and abdomen. The ground color runs from Tawny to Tawny Ochraceous of Ridgway and the black shaft stripes in some specimens are broad and in others practically obsolescent. The color of the under tail coverts varies from tawny to black, heavily vermiculated with white. None of this seems to be attributable to age, for a young male with very small wing spurs, but in fully adult plumage, can have all its color characteristics matched by older specimens with well-developed wing spurs. In fact this young male, while it has a tawny breast and abdomen with practically obsolescent dark shaft streaks, has black flanks and as dark under tail coverts as any of the other specimens examined. Wings 175-183 mm. Females are the same as the darkest-breasted specimens of the forms of leucogenis. The ten specimens examined show no variation in this respect. Wings 155-164 mm. Immatures are similar to those of typical leucogenis, having the fine dusky speckling on the posterior parts of the cheeks and sides of the neck. In this species, however, there seems to be a stronger tendency toward a reddish spotting on the under parts. The four youngest examples examined (primaries not yet fully grown) have the lower neck, chest and upper breast dappled with light rusty. In three 1943 TORRENT DUCKS — CONOVER 355 of these the dappling is very light, in the fourth much heavier. This rusty coloring, strange to say, seems to have no sexual significance, as an older example with the black facial markings of the male well advanced has an even heavier dappling of rusty about the head, neck and chest. Perhaps it is this reddish speckling that is respon- sible for suggestions that the young males go through a female stage with cinnamon-colored under parts before assuming the full adult plumage. In young males the first adult feathers to make their appearance are the black ones of the throat. At about the same time a few of the white-edged scapulars come in, followed by the tertials and the feathers of the mantle. Not until the black head markings are almost com- pleted do the black chest feathers begin to appear, after which come the tawny black-shafted ones of the breast and abdomen. The first adult characters to show in the young females seem to be the dusky and white vermiculations on the posterior cheeks and sides of the neck. The rufous of the under parts first appears on the cheeks, throat, foreneck and under tail coverts and finally on the chest, flanks and the rest of the under parts. This first reddish coloring is just as dark as in the adults; therefore there seems to be no intermediate stage with the under parts light cinnamon-rufous. Downy young are identical with those of colombiana. A correspondent, Mr. Victor Ricke of Lautaro, Cautin, Chile, has sent some information concerning the nesting of this duck. I quote from his account: "Last October I found a nest of Merganetta armata with five eggs here (Lautaro) on the banks of our overflow canal at our hydroelectric plant, a short distance before the water returns to the Cautin River. The nest in question was about one meter above the canal (water) level. While the female sat on the nest, the male kept watch from the opposite bank. This circum- stance led me to find the nest, which was very well hidden in the tall grass. It was necessary to climb down with a rope to reach the nest. The eggs are elongated, similar in color to those of the domestic duck. I took away two eggs first and the female kept on sitting on the nest. When I took away the others, both the female and male abandoned the place. . . . The nest was made with dry grass and had plenty of feathers." From a photograph and a sketch enclosed with the above description I judge the nest to have been situated in a hollow in the almost vertical side wall of the canal about ten feet from the top and three feet above the water. Some fifty feet downstream from the nest, the canal ended, the water falling about 356 FIELD MUSEUM OF NATURAL HISTORY— ZOOLOGY, VOL. 24 thirty feet into the Cautin River. In another letter Mr. Ricke speaks of finding a second nest on the Cautin River on December 1, 1938, containing four eggs, but later speaks of both nests as having five eggs. A female taken at Lautaro, Cautin, December 8, is marked as nesting. In the series of specimens examined is a downy from Coquimbo taken in November and another from Cautin taken in January. There are juveniles with the primaries not fully grown shot in January and March, while older ones just starting to show some adult feathers were collected in February and March. Specimens taken in April, June and July are halfway along toward their final plumage. It seems therefore that these ducks probably attain their adult dress by the time they are a year old. It has been suggested by some that these ducks cannot swim against the current in the swift streams in which they live. While in some places this may be true, I have personally seen a pair swim against a swift current for well over a hundred yards, and then hold themselves stationary in the midst of it, while they watched me as I crouched on the bank. Specimens examined. — CHILE : Guanta, Coquimbo, 1; Balala, Rio Turbio, Coquimbo, 2; Rio Blanco, Santiago, 1; Santiago, 2; Banos de Cauquenes, Rio Claro, 5; Colchagua, 3; Lautaro, Cautin, 10; Villa Rica, Cautin, 2; Casa Richards, Rio Nirehuao, Aysen, 2; unspeci- fied, 5. — ARGENTINA: Las Lajas, Rio Agrio, Neuquen, 1; Bariloche, Rio Negro, 1; Rio Quemquemtreu, Rio Negro, 1; Rio Epuyen, Chubut, 1; Rio Cholila, Chubut, 1. KEY TO MALES OF Merganetta Throat and foreneck white. Mantle brownish gray finely vermiculated with white Af. colombiana Mantle black or black streaked with white M. leucogenis Mantle black (sometimes washed with olive). Under parts not heavily streaked with black M. I. leucogenis Under parts heavily streaked with black M . I. turneri Mantle black streaked with white. Under parts not heavily streaked with black, edges to scapulars buffy gray M. I. garleppi Under parts heavily streaked with black, edges to scapulars buffy white or white M. I. berlepschi Throat and foreneck black, a black stripe from the throat to the eye . . . M. armata UNIVERSITY OF ILLINOIS-URBAN*