Without doubt, waterfowl of the family Anatidae have provided the greatest number and variety of bird hybrids originating from both natural and captive conditions. The recent compilation by Gray (1958) has listed approximately 400 interspecies hybrid combinations in this group, which are far more than have occurred in any other single bird family. Such a remarkable propensity for hybridization in this group provides a great many possibilities for studying the genetics of speciation and the genetics of plum- age and behavior, and it also provides a valuable tool for judging species relationships. It may generally be said that the more closely two species are related the more readily these species will hybridize and the more likely they are to produce fertile offspring. In waterfowl, chromosomal imcompatibility and sterility factors are thought to be in- frequent, a circumstance which would favor the large number of hybrids encountered in this group. In addition to this, however, it can probably be safely concluded that the Anatidae are extremely close-knit in an evolutionary sense, for their behavior, anatomy, and other characteristics all indicate a monophyletic origin. It was for these reamns that Delacour and Mayr (1945), in their revision of the group, sensibly broadened the species, generic, and subfamilial categories, and in so doing greatly clarified natural relationships. Gray's compilation, although it provides an incomparable source of hybrid records, does not attempt to synthesize these data into any kind of biologically meaningful pat- tern. For the past several years I have independently been collecting records and infor- mation on waterfowl hybrids for the purpose of obtaining additional evidence for species relationships and in order better to understand problems of isolating mechanisms under natural conditions. With the publication of Gray's bibliography, little purpose would be served by presenting my compilation in toro, since the two lists are practically identical. However, I believe that condensed graphic summaries, arranged by natural subgroups (tribes) rather than alphabetically as done by Gray, have sufficient value and interest to be presented here. SUMMARY OF INTRATRIBAL HYBRIDS Inasmuch as the great majority of anatid hybrids involve species of the same tribe, and because practically all natural hybrids known are of this kind, they will be dealt with first. Closely related sympatric species which will occasionally hybridize in the wild are most important from the standpoint of isolating mechanisms. The frequency of such hybridization and the fertility of the resulting progeny are of particular importance in this connection. For each tribe in which intratribal hybrids have been recorded a graph (figs. 1-6) has been prepared which indicates the known combinations and also provides an indication of the relative frequency of such hybridization and the degree of hybrid fertility when it is known. Except for the Ringed Teal ("Anas" leucopkrys of Delacour and Mayr, 1945), which is considered here to be congeneric with Amazonetta brasili- ensis of the Cairinini (see Dersheid, 1938), the species constitution of each tribe is that proposed by Delacour and Mayr. However, the exact sequence of species often differs somewhat to conform with what I believe more closely approximates relationships when expressed in a linear fashion. References concerning nearly all the indicated hybrid com- binations can be found in Gray (1958). DENDROCYGNINI e: , INALIS / DENDROCYGNA EYTONI D. BICOLOR D. AUT Vol. 62 CYGNUS ATRATUS C. OLD C. C. COL [o nyBRiDS Fig. 1. Above, intratribal hybrid combinations recorded in the Dendrocygnini. Below, intratribal hybrid combinations recorded in the Anserini. Subfamily Anseranatinae No hybrids involving the monotypic species of this subfamily, the Magpie Goose (Anseranas semipalmata), have been recorded. The species' anatomy deviates markedly from the rest of the group (Miller, 1919), and it is probable that hybridization with other species is genetically impossible. Subfamily Anserinae Dendrocygnini.--Evidence from hybrids within the whistling duck group sheds little light on relationships (fig. 1). It is significant however that no intertribal hybrids in- TADORNINI CHLOEPHAGA MELANOPTER, / / C. POLIOCEPHALA // C. RUB CEPS C. PICTA CYANOCHEN CYANOPTERA iX NEOCHEN dUBATUS / LOPOCHEN AEGYPTIACUS /, / TADORNA TADORNOIDES ß T. VARIEGATA T. RADJAH Fig. 2. lntratribal hybrid combinations recorded in the Tadornini. See figure 1 for meaning of symbols. 1/2IRININI / IBATA / 'EUCOPHRYS / SILIENSIS / IX GALERICULATA PLECTROPTERUS GAMBENSIS CAIRINA MOSCHATA CHENONETTA JU, AMAZONETTA A. BRAS Fig. 3. lntratribal hybrid combinations recorded in the Cairinini. volving this group are known except for a highly dubious record mentioned by Dela- cour (1927) for Dendrocygna viduata x Netta peposaca. This suggests that the tribe is more isolated from the other ariarids than are the geese and swans. Anserini.--The great abundance of swan x goose hybrids known (fig. 1) certainly vindicates the submerging of the previously upheld subfamilies Cygninae and Anserinae, ANAS SIBILATRIX A, AMERICANA A, PENELOPE Fig. 4. Intratribal hybrid combinations recorded in the Anatini (genus Anas only, see text). and the numerous goose hybrids also provide justification for discarding several mono- typic genera such as Eulabeia and Cygnopsis. Even with the broad generic concept of Delacour and Mayr (1945), the several fertile intergeneric crosses between Anser and Branta indicate the very close relationship existing between these two morphologically fairly distinct groups. Subfamily Anatinae Tadornini.--Hybridization in captivity involving species of this tribe is relatively prevalent (fig. 2), probably largely as a result of the active part the female takes in selecting the strongest possible mate, o[ten regardless o[ its species. Since several of the species are only rarely kept in captivity, the potential list of hybrids has undoubtedly not yet been filled. The fertile intergeneric crosses among Alopochen, Tedor,e, Cyano- chcn, and Neochcn 'indicate that the number of genera currently accepted is overly large, and possibly some of these monotypic genera are unrealistic. NETTA RUFIN A. AFFI N. pEPOSACA N. ERYTHROPTHALMA AYTHYA vALLISNERI A A, NYROCA A. AMERICANA A. COLLARIS A. FEEl NA A, INNOTATA A. BAERI A. FULISULA Fig. 5. [ntratribal hybrid combinations recorded in the Aythyini. ß ß ' bin the perching duck group (fig- 3) _ . rising paucity of hy. bnds ,wit_ a . reuresent a collection of carnm.The su __:a;,, hat e mauue  and em --;able by the pu,,j .t primitive characteristics rdafively isdated fores wch share mveral rather to represent the generalized arialine condition from which the oer groups have adap- are known to be fertile, but apparently . . d the Mallard (As fivdy radiad. None of e intraibal hybrids 'ds of the Wo Duck (A spsa) hybrids betweft .... a, robe avvarent asence u, jri basis of gross ohos) are rarely mt[ ......... might be explained on the th the Mdarin Duck (Aix galeculata) - mashina, 1952), but several alleged examples of this cross chromosomal differences ( a ....... ,,bddized th a great many species of Anas, have been rorded. The Wo uc ha= wi mine oi which the Mdarin has also hybridCed, but beharoral (Lore, 1951- 1953) and serdocal (Cotter, 1957) edence dicates that they 1ong in e present tribe. - -ionable rord (e Gray, 1958) o[ a hybrid between the Anatini.A single quest rh nchus and the Gray Duck (Anas New Zeald Bhe Duck (Hymenolaimus malaco Y superciliosa) is the only one which does not involve only those species considered by Delacour and Mayr (1945) to belong to the genus Anas. So many hybrid combinations among species of this genus are known (fig. 4), and such a remarkably high incidence of fertility is present, that one must conclude that the species of dabbling ducks are much more closely related to one another than is often indicated by their external appearance. MERGINI IdELANITTA FUSCA NN, N  M. PERSPICILLATA SOMATERIA IdOLLI$SIMA  M. SPEC TABILI S BUCEI ALA ISLANDICA  B. CLANGULA  X .ERGUS ALBELLUS M. CUCULLATUS / Fig. 6. Intratribal hybrid combinations recorded in the Mergini. Delacour and Mayr's recommendations of a single, inclusive genus rather than the more than twenty which have been proposed is fully justified on this basis. Fertile hybrids of Anas and species of the tribes Tadornini and Aythyini have been occasionally reported, which circumstance illustrates the astonishing capacity for interspec{es fertility .found in this group. Aytkyini.---The high percentage of the potential hybrids among species of diving ducks that are actually realized (fig. 5) forces one to conclude that the Aythyini rep- resent a very closely related group of species, with two being the maximum number of genera that can be reasonably allowed. Indeed, perhaps the best reason for retaining Netta is because it provides such an excellent transitional link between the Anatini and the Aythyini, having provided fertile hybrids with both Anas and Aytkya. The great number of hybrids of Anas x Aythya (five combinations of which have been found to be fertile) and electrophoretic evidence from egg-white proteins (McCabe and Deutsch, 1952) both argue for a very close relationship between the groups. Mergini.--The relatively few hybrid records for the sea duck group (fig. 6) very likely reflects their relative infrequence in captive collections, the difficulties of breeding them, and their secondary importance as sporting birds, which reduces the number of wild-taken hybrids. Except for the remarkable intertribal hybrid between Anas castanea and Clangula kyernalis, which Delacour (1956) bred at Cleres, and an equally curious hybrid between Aix galericulata and Clangula byemalls from Tring (specimen in the American Museum of Natural History), most of the hybrids represent wild-taken speci- mens. The recent evidence (Humphrey, 1958) that eiders should be placed in a sep- arate tribe near the dabbling ducks might be supported by two alleged crosses of Anas x Somateria, but behavioral evidence (D. F. McKinney, unpublished MS) argues strongly against this. Table 1 Intratribal and intertribal hybrid combinations recorded in the Anatidae. Numbers indicate total interspecific hybrid combinations recorded within and among tribes. Reciprocal crosses (such as Mallard x Pintail and Pintail x Mallard) are considered as single combinations. AnranaLini .................................... Dendrocygnini .... $ ............................ Anserini ........ 65 15 5 3 ............ Tadornini ........ 15 21 8 5 1 1 .... Cairinini ........ 5 8 5 28 8 2 .... Anatini ........ 3 5 28 115 45 7 .... Aythyini ............ 1 8 45 35 6 .... Mergini ............ 1 2 7 6 9 .... Oxyurini ................................ 1 Oxyurini.--Except for a highly questionable record of a hybrid between a Ruddy Duck (Oxyura ]arnaicensis) and a Greater Scaup (Aythya rnarila) mentioned by Sibley (1938), there are no positive records of interspecific hybridization involving any species of the stiff-tailed ducks. Bond (1950) has indicated that intermediate forms between Oxyura jarnaicensis andina and O. ]erruginea occur in Colombia, but these probably should be considered subspecific intergrades rather than hybrids (Johnsgard, unpub- lished MS). One possible hybrid between ]erruginea and Oxyura vittata has also been mentioned (Hellrn3yr , 1932), but the evidence for this cross is still inadequate. SUMMARY OF INTERGENERIC AND INTERTRIBAL HYBRIDS A compilation of records of intergeneric and intertribal hybrids provides a means of estimating the degree of genetic isolation existing among these higher categories. As such it helps to establish a sequence of categories that most closely reflects actual rela- tionships. Thus by totalling the number of known intergeneric hybrid combinations (not total hybrids), a rough index is obtained to the degree of genetic difference among various genera (fig. 7). Such a compilation clearly indicates the great latitude of hybrid potentiality present in the group (very questionable records are not included), particu- larly in the subfamily Anatinae. Thus, for example, the Mallard (Anas platyrkynchos) has been hybridized with both the Gray-lag Goose (Anser anser) and the Common Mer- ganser (Merganser merganser), two species which practically encompass the greatest range of morphological variation encountered in the entire family Anatidae. If a compilation of hybrid records is made at the tribal level (intertribal and intra- tribal combinations), as has been done in table 1, a further fact becomes clear. This is that the tribe Cairinini belongs between the Tadornini and the Anatini, rather than between the Aythyini and the Mergini, as originally suggested by Delacour and Mayr (1945). In addition, the table illustrates the fact that the whistling ducks (Dendro- cygnini) are a distinctly isolated group, whereas the geese and swans (Anserini) are not  MERGINI TADORNINI n. CAIRININI ANSERINI CYGNUS ANSER TADO!  --7 or more Fig. 7. Intratribal and intertribal hybrid combinations recorded among the genera of the Anatidae. Numbers indicated in key refer to total intergeneric combinations recorded. markedly set off from the Anatinae. Likewise the stiff-tail group (Oxyurini) appears to be rather isolated from the other tribes. However, the rather obvious overall conclusion to be drawn from all these facts is that broad nomenclatural categories are greatly pref- erable to narrow ones in the Anatidae and that they help to emphasize actual biological relationships far better than does the highly "split" taxonomic hierarchy which is still in general usage as in the fifth edition of the American Ornithologists' Union Check-list. CONCLUSIONS A compilation of published and unpublished records of hybridization in the family Anatidae indicates that the genera and tribes proposed by Delacour and Mayr (1945) are well grounded. With but a single exception, that of the Cairinini, fertile hybrids resulting from intrageneric crosses and intratribal hybrids are more frequent than are intertribal combinations. The exceptional situation found in the Cairinini is largely the result of the Wood Duck (Aix sponsa), which has hybridized with 16 species of Anas but with only two species of its own tribe. This and other evidence clearly indicates that the correct position of the Cairinini is between the Tadornini and the Anatini, rather than between the Aythyini and the Mergini. Furthermore, most of the lumping of genera, especially in the Anserini, Anatini and Aythyini, as proposed by Delacour and Mayr, is well justified by the hybrid records. ACKNOWLEDGMENTS Financial support in the form of fellowships was received from the Graduate School of Cornell University and the National Science Foundation in the course of the period while the data presented here were being assembled. Dr. Charles G. Sibley read and commented on an early version of the manuscript. LITERATURE CITED Bond, J. 195o. Check-list of birds of the West Indies (Acad. Nat. Sci. Philadelphia). Cotter, W. B., Jr. 1957. A serological analysis of some anatid classifications. Wfison Bull., 69:291-300. Delacour, J. 1927. Notes sur quelques hybrides. L'Oiseaux, 8:276-284. 1956. The waterfowl of the world. Vol. II (Country Life, London). Delacour, J., and Mayr, E. 1945. The family Anatidae. Wilson Bull., 57:3-55. Dersheid, J. M. 1938. [A note on the systematic position of Amazonefta and Calonetta.] Bull. Brit. Ornith. Club, 58: 62-63. Gray, A. P. 1958. Bird hybrids. Tech. Comm. No. 13, Comm. Bur. Animal Breeding and Genetics, Edin- burgh. Hellmayr, C. E. 1932. The birds of Chile. Field Mus. Nat. Hist. Publ. 380, Zool. Ser., 19:1-472. Humphrey, P. S. 1958. Classification and systematic position of the eiders. Condor, 60:129-135. Lorenz, K. Z. 1951-1953. Comparative studies on the behavior of the Anatinae. Avic. Mag., 57:157-182; 58:8-17, 61-72, 86-94, 172-184; 59:24-34, 80-91. McCabe, R. A., and Deutsch, H. F. 1952. The relationships of certain birds as indicated by their egg white proteins. Auk, 91:1-18. Miller, W. De W. 1919. Notes on the structure of Anserarias semipalmata. Auk, 36:562-564. Sibley, C. L. 1938. Hybrids of and with North American Anatidae. Proc. 9th Internat. Ornith. Congr., Rouen, 1938:327-335. Yamashina, Y. 1952. Classification of the Anatidae based on the cytogenetics. Papers from the Coordinating Committee for research in Genetics, 3:1-24. Department o! Conservation, Cornell University, Ithaca, New York, July 16, 1959.