Chapter II: Part 2

In _Passer_, _Estrilda_ and _Poephila_, and in all the cardueline finches examined, the bellies of the _m. flexor perforans et perforatus digiti II_ and the _m. flexor perforans et perforatus digiti III_ are more intimately connected than they are in the other species studied. Thus, the amount of independent action of these muscles in _Passer_, in the estrildines, and in the carduelines probably is reduced.

In _Passer_, the estrildines, and the carduelines the edges of the sheathlike tendon of insertion of the _m. perforatus digiti III_ are thickened; as a result the insertion appears superficially to be double but closer examination reveals that there is a fascia stretched between the thickened edges. In the other species examined, the insertion is sheathlike throughout and there are no thick areas. I cannot explain this on the basis of function. The difference, however, is obvious and constant.

Aside from the differences noted above, there were variations of muscle pattern that seem to be significant only in _Vireo olivaceus_. In this species the central, aponeurotic portion of the _m. iliotibialis_ is absent. The origin of the _m. adductor longus et brevis_ is from the dorsal edge of the ischiopubic fenestra and not from the membrane covering this fenestra. The origin of the _pars posticus_ of this muscle, furthermore, is fleshy and not tendinous as it is in the other species. The _m. flexor perforatus digiti II_ is larger and more deeply situated in _Vireo_ and has, furthermore, no connection with the _m. flexor hallucis longus_. The latter muscle is smaller and weaker than in any of the other species and has only one (the posterior) head of origin. The _m. flexor hallucis brevis_, on the contrary, is larger than in the other birds, compensating, probably, for the small _m. flexor hallucis longus_. In those differences, however, which separate the carduelines and ploceids from the other birds studied, _Vireo_ resembles, in every instance, the richmondenines, emberizines, tanagers, warblers, and blackbirds.

On the basis of differences in leg-musculature the species which are now included in the Family Fringillidae may be separated into two groups. One group includes the richmondenines and the emberizines; the other, the carduelines. The muscle patterns of the legs of the birds of the first group are indistinguishable from those of _Seiurus_, _Icterus_, _Molothrus_, and _Piranga_, and except for the differences noted are similar to those in _Vireo_. The carduelines, on the other hand, are similar in every point of leg-musculature to the ploceids which were studied. Thus, the heterogeneity of the Family Fringillidae, as now recognized, is emphasized by differences in the muscle patterns of the leg.

COMPARATIVE SEROLOGY

General Statement

The application of serological techniques to the problems of animal relationships has been attempted with varying degrees of success over a period of approximately fifty years. Few of the earlier studies were of a quantitative nature, but within the past decade, satisfactory quantitative serological techniques have been developed whereby taxonomic relationships may be estimated. The usefulness of comparative serology in taxonomy has been demonstrated in investigations of many groups wherein results obtained have, in most instances, been compatible with the results obtained by more conventional methods, such as comparative morphology. As Boyden (1942:141) stated, "comparative serology ... is no simple guide to animal relationship." However, the objectiveness of its methods, the fact that it has its basis in the comparisons of biochemical systems which seem to be relatively slow to change in response to external environmental influences, and the fact that the results are of quantitative nature favor, where possible, the inclusion of data from comparative serology along with that from more conventional sources when an attempt is made to determine the relationships of groups of animals.

The application of serological methods in ornithology has not been extensive. Irwin and Cole (1936) and Cumley and Irwin (1941, 1944) used two species of doves and their hybrids and demonstrated that a distinction between the red cells of these birds could be made by use of immunological methods involving the agglutinin reaction. McGibbon (1945) was able to distinguish the red cells of interspecific hybrids in ducks by similar methods. Irwin (1953) used similar techniques in his study of the evolutionary patterns of some antigenic substances of the blood cells of birds of the Family Columbidae. Sasaki (1928) demonstrated the usefulness of the precipitin technique in distinguishing species of ducks and their hybrids. This technique was used successfully also by DeFalco (1942) and by Martin and Leone (1952). Working with groups of known relationships, these investigators showed that the "accepted" systematic positions of certain birds were confirmed by serological procedures. The precipitin reaction, however, has never been applied to actual problems in avian taxonomy prior to the present study.

Preparation of Antigens

Although most previous work in comparative serology in which precipitin tests were used has involved the use of whole sera as antigens, Martin and Leone (1952) indicated that tissue extracts are satisfactory as antigens and that serological differentiation can be obtained with these extracts and the antisera to them. I decided, therefore, to use such extracts in these investigations, since the small sizes of the birds to be tested made it impracticable to obtain enough whole sera.

Most of the birds used were obtained by shooting, but a few were trapped and the exotic species were purchased alive from a pet dealer. When a bird was killed, the entire digestive tract was carefully removed to prevent the escape of digestive enzymes into the tissues and to prevent putrefaction by action of intestinal bacteria. As soon as possible (and within three hours in every instance) the bird was skinned, the head, wings, and legs were removed, and the body was frozen. Each specimen, consisting of trunk, heart, lungs, and kidneys, was wrapped separately and carefully in aluminum foil to prevent dehydration of the tissues. The specimens were kept frozen until the time when the extracts were made.

When an extract was to be prepared, the specimen was allowed to thaw but not to become warm. In the cold room with the temperature of all equipment and reagents at 2 deg.C., the specimen was placed in a Waring blender with 0.9 per cent aqueous solution of NaCl buffered with M/150 K_{2}HPO_{4} and M/150 Na_{2}HPO_{4} to a pH of 7.0. The amount of reagent used was 75 ml. of saline for each gram of tissue to be extracted. The tissues were minced in the blender, allowed to stand at 2 deg.C. for 72 hours, and the tissue residues removed by centrifugation in a refrigerated centrifuge. Formalin was added to a portion of the supernatant in the amount necessary to make the final dilution 0.4 per cent. This formolization was found to be necessary to inhibit the action of autolytic enzymes over the period of time required to complete the investigations. The effects of formolization on the antigenicity and reactivity of proteins are discussed later. It was necessary to sterilize and clarify the "native" (unformolized) extracts; this was done by filtration through a Seitz filter. These "native" substances were used only in the early stages of the investigation (see below). The filtrate was bottled and stored at 2 deg.C. In the early stages of this investigation clarification of the formolized extract was accomplished by the same sort of filtration. It was determined, however, that centrifugation in a refrigerated centrifuge at high speeds (17,000g) served the same purpose and was quicker. The formolized extracts were bottled and also stored at 2 deg.C. (although refrigerated storage of the formolized extracts does not seem necessary). For each extract the amount of protein present was determined colorimetrically by the method of Greenberg (1929) with a Leitz Photrometer.

Species for which extracts were prepared and the protein values of the extracts are listed in Table 1. Extracts of some species were used throughout most of the experiment; extracts of others were used only when needed for purposes of comparison.

TABLE 1.--Species from Which Extracts Were Prepared and Injection
Schedules for Extracts Against Which Antisera Were Produced

==========================+==========+================================= | Protein, | SPECIES | gms. per | Injection schedules for | 100 ml. | production of antisera --------------------------+----------+--------------------------------- _Myiarchus crinitus_ | 0.65 | Series 1: Intravenous, 0.5, 1.0, (Linnaeus) | | 2.0, and 4.0 ml. --------------------------+----------+--------------------------------- _Passer domesticus_ | 1.40 | Series 1: Subcutaneous, 0.5, | | 1.0, 2.0, and 4.0 ml. --------------------------+----------+--------------------------------- _Estrilda amandava_ | 0.45 | [A]Series 1: Intravenous, 0.5, | | 1.0, 2.0, and 4.0 ml. | | | | [A]Series 2: Subcutaneous, 0.5, | | 1.0, and 2.0 ml. | | | | Intraperitoneal, 8.0 ml. --------------------------+----------+--------------------------------- _Poephila guttata_ | 0.56 | [A]Same as for _Estrilda_. --------------------------+----------+--------------------------------- _Molothrus ater_ | 0.65 | Series 1: Intravenous and | | subcutaneous, respectively, 0.5 | | and 0.5 ml., 1.0 and 1.0 ml., | | 3.0 and 1.0 ml., 5.0 and 3.0 ml. | | | | Series 2: Subcutaneous, 0.5, | | 1.0, 2.0 and 4.0 ml. --------------------------+----------+--------------------------------- _Piranga rubra_ | 0.50 | Same as for _Molothrus_. --------------------------+----------+--------------------------------- _Richmondena cardinalis_ | 0.70 | [A]Same as for _Estrilda_. --------------------------+----------+--------------------------------- _Richmondena cardinalis_ | 0.60 | Same as for _Spinus_. --------------------------+----------+--------------------------------- _Passerina cyanea_ | 0.45 | Antiserum not prepared. --------------------------+----------+--------------------------------- _Spiza americana_ | 0.70 | Same as for _Molothrus_. --------------------------+----------+--------------------------------- _Carpodacus purpureus_ | 0.50 | Antiserum not prepared. --------------------------+----------+--------------------------------- _Spinus tristis_ | 0.49 | Series 1: Intravenous, 0.5, 1.0, | | 2.0, and 4.0 ml. | | | | Series 2: Intravenous, 0.5, 1.0, | | 2.0, and 4.0 ml. | | | | Series 3: Subcutaneous, 0.5, | | 1.0, 2.0, and 4.0 ml. --------------------------+----------+--------------------------------- _Pipilo erythrophthalmus_ | 0.92 | Antiserum not prepared. --------------------------+----------+--------------------------------- _Junco hyemalis_ | 0.56 | Same as for _Spinus_. --------------------------+----------+--------------------------------- _Spizella arborea_ | 0.48 | Same as for _Spinus_. --------------------------+----------+--------------------------------- _Zonotrichia querula_ | 0.48 | Same as for _Spinus_. --------------------------+----------+--------------------------------- _Zonotrichia albicollis_ | 0.92 | Antiserum not prepared. (Gmelin) | | --------------------------+----------+---------------------------------

[A] Antiserum prepared against formolized antigen.

Preparation of Antisera

All antisera were produced in rabbits (laboratory stock of _Oryctolagus cuniculus_). Three methods of injection of antigen were used in various combinations: intravenous, subcutaneous, and intraperitoneal. Injection schedules used in the production of each antiserum are listed in Table 1. Both formolized and "native" antigens were used. Each rabbit received one or more series of four injections, each injection being administered on alternate days and doubling in amount: 0.5 ml., 1.0 ml., 2.0 ml., and 4.0 ml. In all but two instances more than one series of injections was necessary to produce a useful antiserum. More than two series, however, resulted in little or no improvement of the reactivity of the antiserum.

The injection-series were separated by intervals of eight days. On the eighth day after the last injection of each series, 10 ml. of blood were withdrawn from the main artery of the ear of the rabbit, and the antiserum was used in a homologous precipitin test to determine its usefulness. If the antiserum contained sufficient amounts of antibodies to conduct the projected tests, the rabbit was completely exsanguinated by cardiac puncture, by using an 18-gauge needle and a 50 ml. syringe. The whole blood was placed in clean test tubes and allowed to clot. It was allowed to stand at 2 deg.C. for 12 to 18 hours so that most of the serum would be expressed from the clot. The serum was then decanted, centrifuged to remove all blood cells, sterilized in a Seitz filter, bottled in sterile vials, and stored at 2 deg.C. until used.

Methods of Serological Testing

The precipitin reaction is the most successful of the serological techniques thus far devised for systematic comparisons. The reaction occurs because antigenic substances introduced into the body of an animal cause the formation of antibodies which precipitate antigens when the two are mixed. The antisera which are produced show quantitative specificities in their actions; therefore, when an antiserum containing precipitins is mixed with each of several antigens, the reaction involving the homologous antigen (that used in the production of the antiserum) is greater than those reactions involving the heterologous antigens (antigens other than those used in the production of the antiserum). Furthermore, the magnitudes of the reactions between the antiserum and the heterologous antigens vary according to the degrees of similarity of these antigens to the homologous one.

The method of precipitin testing follows that outlined by Leone (1949). The Libby (1938) Photronreflectometer was used to measure the turbidities developed by the interaction of antigen and antiserum. With this instrument parallel rays of light are passed through the turbid systems being measured. Light rays are reflected from the suspended particles to the sensitive plate of a photoelectric cell; this generates a current of electricity which causes a deflection on a galvanometer. The deflection is proportional to the amount of turbidity developed and readings may be taken directly from the scale of the instrument.

The reaction-cells of the photronreflectometer are designed to operate with a volume of 2 ml.; therefore, this volume was used in all testing. In every series of tests the amount of antiserum was held constant and the amount of antigen was varied. The volume for each antigen dilution was always 1.7 ml., and to this was added 0.3 ml. of antiserum to make up a volume of 2 ml.

TABLE 2.--Percentage values obtained from analyses of precipitin
reactions. Numerals represent relative amounts of reaction between
antigens and antisera. Homologous reactions are arbitrarily valued
as 100 per cent, and heterologous reactions are expressed
accordingly. _Comparisons are meaningful only if made within each
horizontal row of values._

Table headings:
Col A: _Estrilda amandava_
Col B: _Poephila guttata_
Col C: _Piranga rubra_
Col D: _Richmondena cardinalis_
Col E: _Spiza americana_
Col F: _Spinus tristis_
Col G: _Junco hyemalis_
Col H: _Zonotrichia querula_

========================+============================================== | ANTISERA ANTIGENS +-----+-----+-----+-----+-----+-----+-----+---- | A | B | C | D | E | F | G | H ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Passer domesticus_ | 75 | 74 | 73 | 66 | 81 | 72 | ... | 81 ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Estrilda amandava_ | 100 | 88 | 75 | ... | 79 | 72 | 53 | ... ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Poephila guttata_ | 95 | 100 | 77 | 67 | 87 | 81 | ... | ... ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Molothrus ater_ | 66 | 54 | 69 | 65 | 86 | 75 | 69 | 75 ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Piranga rubra_ | ... | ... | 100 | ... | ... | ... | ... | 89 ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Richmondena cardinalis_| 75 | 80 | 91 | 100 | 98 | 65 | 88 | 91 ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Spiza americana_ | 65 | 68 | ... | 71 | 100 | 64 | 67 | 80 ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Carpodacus purpureus_ | 70 | 71 | 71 | 61 | 89 | 93 | 53 | 70 ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Spinus tristis_ | 72 | 74 | 73 | 60 | 89 | 100 | 60 | ... ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Junco hyemalis_ | 64 | 56 | 74 | 65 | 87 | 68 | 100 | ... ------------------------+-----+-----+-----+-----+-----+-----+-----+---- _Zonotrichia querula_ | 65 | 71 | ... | 67 | 89 | 75 | ... | 100 ------------------------+-----+-----+-----+-----+-----+-----+-----+----

Antigens were diluted with 0.9 per cent phosphate-buffered saline solution. Tests were run in standard Kolmer test-tube racks, each test consisting of 12 tubes. Each dilution was made on the basis of the known protein concentration of the antigen. The first tube contained an initial dilution of 1 part protein in 250 parts saline and each successive tube contained a protein dilution one-half the concentration of the preceding tube, ranging up to 1:512,000. Saline controls, antiserum controls, and antigen controls were maintained with each test to determine the turbidities inherent in these solutions. These control-turbidities were deducted from the total turbidity developed in each reaction-tube, the resultant turbidity then being considered as that which was caused by the interaction of antigens and antibodies. The turbidities were allowed to develop over a 24-hour period. In the early stages of this investigation the reactions were allowed to take place at 2 deg.C. in order to inhibit bacterial growth.

Later tests were carried out at room temperatures, and bacterial growth was prevented by the addition to each tube of 'Merthiolate' in a final dilution of 1:10,000.

Experimental Data

Corrected values for the turbidities obtained were plotted with the turbidity values on the ordinate and the antigen dilutions on the abscissa. The homologous reaction was the standard of reference for all other test reactions with the same antiserum. By summing the plotted turbidity readings, numerical values are obtained which are indices serving to characterize the curves. Such values were converted to percentage values, that of the homologous reaction being considered 100 per cent. These values, plus the curves, provide the data by means of which the proteins of the birds may be compared. Plots representative of the precipitin curves are presented in Figs. 10 to 21. For convenience each plot represents only several of the 10 curves obtained with each antiserum.

A summary of the serological relationships of the birds involved in the precipitin tests is presented in Table 2, in which percentage values are presented. Since the techniques involved in testing were greatly improved as the investigation proceeded, the summary is based solely on those tests run in the later stages of the investigation. For reasons which will become apparent in later discussion, it should be emphasized that in Table 2 comparisons may be made only within each horizontal row of values.

Discussion of the Serological Investigations

One of the problems met early in this investigation was instability of the proteins in the extracts that were prepared. Extracts in which no attempt was made to inactivate the enzymes present proved unsatisfactory. It was necessary to maintain the temperature of the "native" antigens at 2 deg.C, and all work with such antigens had to be performed at this temperature. This arrangement was inconvenient; furthermore, inactivation of the enzymes was not complete even at this low temperature, and some denaturation of the proteins took place as evidenced by the gradual appearance of insoluble precipitates in the stored vials.

The preservatives, 'Merthiolate' and formalin, were used in an attempt to inhibit the autolytic action of the enzymes present. Formalin, when added to make a final dilution of 0.4 per cent, proved to be the more satisfactory of the two preservatives and was used throughout most of the work. Formalin caused slight denaturation of some of the proteins, but this effect was complete within a few hours, after which any denatured material was removed by filtration or centrifugation. The proteins remaining in solution were stable over the period necessary to complete the investigations.

The addition of formalin reduces the reactivity of the extracts when they are tested with antisera prepared against "native" antigens and causes changes in the nature of the precipitin curves. This effect has been pointed out by Horsfall (1934) and by Leone (1953) in their work on the effects of formaldehyde on serum proteins. Their data indicate, however, that even though changes in the immunological characteristics of proteins are brought about by formolization, the proteins retain enough of their specific chemical characteristics to allow consistent differentiation of species by immunological methods. In the tests which I performed, the relative positions of the precipitin curves, whether native or formolized extracts were involved, remained unchanged (Figs. 10, 11). _All data used in interpretation of the serological relationships were obtained from tests in which formolized antigens of equivalent age were used._

Only three antisera were produced against formolized antigens, all others being produced against "native" extracts. The formolized antigens seemed to have a greater antigenicity, in most instances, than did those which were unformolized, and precipitin reactions involving antisera produced against formolized antigens developed higher turbidities. The antisera produced against formolized antigens were equal to but no better than those prepared against "native" extracts in separating the birds tested (Figs. 12, 13).

The rabbit is a variable to be considered in serological tests. Two rabbits exposed to the same antigen, under the same conditions, may produce antisera which differ greatly in their capacities to distinguish different antigens. It is logical to assume, therefore, that two rabbits exposed to different antigens may produce antisera which also differ in this respect. This explains the unequal values of reciprocal tests shown in Table 2. Thus, in the test involving the antiserum to the extracts of _Richmondena_, a value of 71 per cent was obtained for _Spiza_ antigen, whereas in the test involving anti-_Spiza_ serum, a value of 98 per cent was obtained for _Richmondena_ antigen. In Table 2, therefore, comparisons may be made only among values for the proteins of birds tested with the same antiserum.

Since the amount of any one antiserum is limited, there is, of necessity, a limit as to the number of birds used in a series of serological tests. Therefore, although the results reveal the actual serological relationships of the individual species, interpretation of the relationships of the taxonomic groups must be undertaken with the realization that such an interpretation is based on tests involving relatively few species of each group. It is reasonable to assume, however, that a species which has been placed in a group on the basis of resemblances other than serological resemblance would show greater serological correspondence to other members of that group than it would to members of other groups. Specifically, in the Fringillidae and their allies, there seems to be little reason to doubt that genera, and even subfamilies, are natural groups. This is illustrated in tests involving closely related genera: _Richmondena_ and _Spiza_ (Figs. 14, 15, 18), _Estrilda_ and _Poephila_ (Fig. 21), _Spinus_ and _Carpodacus_ (Figs. 12, 17, 19, 20). In each of these tests the pairs of genera mentioned show greater serological correspondence to each other than they do to other kinds involved. This point is illustrated further by a test (not illustrated) involving _Zonotrichia querula_ (the homologous antigen) and _Zonotrichia albicollis_. Although this test was one of an earlier series in which difficulties were encountered (the data, therefore, were not used), it is of interest that the two species were almost indistinguishable serologically.

The serological homogeneity of passeriform birds is emphasized by the fact that the value of every heterologous reaction was more than 50 per cent of the value of the homologous reaction, except in the test involving the anti-_Richmondena_ serum and _Myiarchus_ (Fig. 13) in which the value of the heterologous reaction was 45 per cent. Because most ornithologists consider these genera to be only distantly related (they are in different suborders within the Order Passeriformes), the relatively high value of the heterologous reaction emphasizes the close serological correspondence of passerine birds and indicates that small consistent serological differences among these birds are actually significant. The possibility that some of the serological correspondence is due to the "homologizing" effect of formalin on proteins should not be excluded. I think, however, that this effect is not entirely responsible for the close correspondence observed here.

An additional point to consider in interpretation of the serological tests is that the techniques used tend to separate sharply species that are closely related whereas species that are distantly related are not so easily separated. In other words, comparative serological studies with the photronreflectometer tend to minimize the differences between distant relatives and to exaggerate the differences between close relatives.

In analyzing the serological relationships of the species used in this study, it becomes obvious that two or more series of tests must be considered before the birds can be placed in relation to each other. For example, the data presented in Fig. 14 indicate that _Spiza_ and _Molothrus_ show approximately the same degree of serological correspondence to _Richmondena_. This does not imply necessarily that _Spiza_ and _Molothrus_ are closely related. If Fig. 15 is examined, it can be determined that _Richmondena_ shows much greater serological correspondence to _Spiza_ than does _Molothrus_. Thus, an analysis of both figures serves to clarify the true serological relationships of the three genera. By reference to other series of tests involving these three birds a more exact determination of their relationships may be obtained.

To illustrate this point by a hypothetical example, two species might seem equidistant, serologically, from a third species. Additional testing should indicate if the first two species are equidistant in the same direction (therefore, by implication, close relatives) or in opposite directions (therefore, distant relatives). A single test supplies only two dimensions of a three dimensional arrangement.

It is impossible to interpret and to picture the serological data satisfactorily in two dimensions; therefore, a three-dimensional model (Figs. 22, 23) was constructed to summarize the serological relationships of the birds involved. Each of the eleven kinds used consistently throughout the investigation is represented in the model. By use of the percentage values (Table 2), each bird was located in relation to the other birds. Where possible, averages of reciprocal tests (Table 3) were used in determining distances between the elements of the model. In this way seven of the birds were accurately located in relation to each other. Lacking reciprocal tests, the positions of the other birds were determined by the values of single tests (Table 4). Although these birds were placed with less certainty, at least four points of reference were used in locating each species. At least one serological test is represented by each connecting bar in the model. The lengths of the bars connecting any two elements were determined as follows: a percentage value (Table 3 and Table 4) representing the degree of serological correspondence between two birds was subtracted from 100 per cent; the remainder was multiplied by a factor of five to increase the size of the model and the product was expressed in millimeters; a bar of proper length connects the two elements involved.

From the model it is observed that, _Molothrus_ and _Passer_ excluded, the birds fall into two distinct groups: one includes _Piranga_, _Richmondena_, _Spiza_, _Junco_, and _Zonotrichia_; the other includes _Estrilda_, _Poephila_, _Carpodacus_, and _Spinus_.

TABLE 3.--Reciprocal Values Used to Determine Distances Between
Elements of the Model; Each Value Represents the Average of
Serological Tests Between the Species Involved

Table Headings:
Col A: _Estrilda amandava_
Col B: _Poephila guttata_
Col C: _Richmondena cardinalis_
Col D: _Spiza americana_
Col E: _Spinus tristis_
Col F: _Junco hyemalis_
Col G: _Zonotrichia querula_

==========================+====+====+====+====+====+====+====+ | A | B | C | D | E | F | G | --------------------------+----+----+----+----+----+----+----+ _Estrilda amandava_ | .. | 92 | .. | 72 | 72 | 59 | .. | --------------------------+----+----+----+----+----+----+----+ _Poephila guttata_ | 92 | .. | 74 | 78 | 78 | .. | .. | --------------------------+----+----+----+----+----+----+----+ _Richmondena cardinalis_ | .. | 74 | .. | 85 | 63 | 77 | 79 | --------------------------+----+----+----+----+----+----+----+ _Spiza americana_ | 72 | 78 | 85 | .. | 77 | 77 | 85 | --------------------------+----+----+----+----+----+----+----+ _Spinus tristis_ | 72 | 78 | 63 | 77 | .. | .. | .. | --------------------------+----+----+----+----+----+----+----+ _Junco hyemalis_ | .. | .. | 77 | 77 | .. | .. | .. | --------------------------+----+----+----+----+----+----+----+ _Zonotrichia querula_ | .. | .. | 79 | 85 | .. | .. | .. | --------------------------+----+----+----+----+----+----+----+

TABLE 4.--Single Values Used to Determine Distances Between Elements
of the Model; Each Value Represents a Single Test Between the
Species Involved

Table headings:
Col A: _Estrilda amandava_
Col B: _Poephila guttata_
Col C: _Piranga rubra_
Col D: _Richmondena cardinalis_
Col E: _Spinus tristis_
Col F: _Junco hyemalis_
Col G: _Zonotrichia querula_

==========================+====+====+====+====+====+====+====+ | A | B | C | D | E | F | G | --------------------------+----+----+----+----+----+----+----+ _Passer domesticus_ | .. | 74 | 73 | .. | 72 | .. | .. | --------------------------+----+----+----+----+----+----+----+ _Molothrus ater_ | .. | 54 | .. | 65 | .. | 69 | 75 | --------------------------+----+----+----+----+----+----+----+ _Piranga rubra_ | .. | 77 | .. | 91 | 73 | 74 | .. | --------------------------+----+----+----+----+----+----+----+ _Carpodacus purpureus_ | 70 | 71 | .. | 61 | 93 | .. | .. | --------------------------+----+----+----+----+----+----+----+

FIG. 10. Reactions of unformolized antigens of _Richmondena_,
_Zonotrichia_, and _Molothrus_ with anti-_Richmondena_ serum.
FIG. 11. Reactions of formolized antigens of _Richmondena_,
_Zonotrichia_, and _Molothrus_ with anti-_Richmondena_ serum.
FIG. 12. Reactions of anti-_Richmondena_ serum prepared against
native antigen with antigens of _Richmondena_, _Zonotrichia_,
_Carpodacus_, and _Spinus_.
FIG. 13. Reactions of anti-_Richmondena_ serum prepared against
formolized antigen with antigens of _Richmondena_, _Zonotrichia_,
_Poephila_, _Spinus_, and _Myiarchus_.]

FIG. 14. Serological relationships of _Richmondena_, _Spiza_, and
_Molothrus_.
FIG. 15. Serological relationships of _Richmondena_, _Spiza_, and
_Molothrus_.
FIG. 16. Serological relationships of _Carpodacus_ with the
richmondenine-emberizine-thraupid assemblage.
FIG. 17. Serological relationships of _Carpodacus_ and _Spinus_ with
_Richmondena_ and _Junco_.]

FIG. 18. Serological relationships of _Spinus_ and _Poephila_ with
the richmondenines.
FIG. 19. Serological relationships of _Carpodacus_ and _Spinus_
with _Richmondena_ and _Piranga_.
FIG. 20. Serological relationships of _Poephila_ and Richmondena
with the carduelines.
FIG. 21. Serological relationships of _Richmondena_ and _Spinus_
with the estrildines.]

Genera Pi . . . . _Piranga_
C . . . . _Carpodacus_ Po . . . . _Poephila_
E . . . . _Estrilda_ R . . . . _Richmondena_
J . . . . _Junco_ Sn . . . . _Spinus_
M . . . . _Molothrus_ Sz . . . . _Spiza_
Pa . . . . _Passer_ Z . . . . _Zonotrichia_]

Genera Pi . . . . _Piranga_
C . . . . _Carpodacus_ Po . . . . _Poephila_
E . . . . _Estrilda_ R . . . . _Richmondena_
J . . . . _Junco_ Sn . . . . _Spinus_
M . . . . _Molothrus_ Sz . . . . _Spiza_
Pa . . . . _Passer_ Z . . . . _Zonotrichia_]

Within the richmondenine-emberizine-thraupid assemblage, _Junco_ and _Zonotrichia_ constitute a sub-group apart from the others. _Piranga_ and _Richmondena_ show close serological correspondence. The present taxonomic position of _Spiza_ in the Richmondeninae, which has been questioned by Beecher (1951a:431; 1953:309), is corroborated at least insofar as the serological evidence is concerned. Certainly, serological correspondence of _Spiza_ with the richmondenine-emberizine-thraupid assemblage is greater than with any other group of birds tested.

It is obvious that the serological affinities of the carduelines do not lie with the richmondenines, emberizines, or thraupids. The carduelines show greater serological correspondence with the estrildines than they do with any of the other groups tested. Further serological investigation involving other species, however, is necessary before the nearest relatives of the carduelines can be determined with certainty.

The two estrildines tested (_Estrilda_ and _Poephila_) show close serological relationship. Their nearest relatives, serologically, seem to be the carduelines. The classification (Wetmore, 1951) that places _Passer_ in the same family with the estrildines is not upheld by the serological data available. _Passer_ is not, serologically, closely related to any of the birds tested. It is of interest that Beecher (1953:303-305), on the basis of jaw musculature, places _Passer_ and the estrildines in separate families (Ploceidae and Estrildidae, respectively).

_Molothrus_ shows greater serological correspondence to the richmondenine-emberizine-thraupid assemblage than to any of the other birds tested. It is definitely set apart from this group, however, and its position, serologically, is compatible with that based on evidence from other sources.

There seems to be but little argument among ornithologists that icterids, fringillids, and ploceids constitute families which are distinct from one another. If, then, the serological differences between _Molothrus_ (Icteridae) and _Richmondena_ (Fringillidae), between _Molothrus_ and _Zonotrichia_ (Fringillidae), and between _Richmondena_ and _Poephila_ (Ploceidae) are indicative of family differences, there are four families represented by the birds involved. _Molothrus_ represents one family; _Piranga_, _Richmondena_, _Spiza_, _Junco_, and _Zonotrichia_, a second; _Estrilda_, _Poephila_, _Carpodacus_, and _Spinus_, a third; and _Passer_, a fourth.

CONCLUSIONS

The heterogeneity of the Family Fringillidae has been emphasized by many authors. The relationships of the species now included in this Family have been the subject of much discussion and constitute an important problem in avian systematics.

Sushkin's studies (1924, 1925) of features of the horny and bony palates have served as a basis for the present division of the Family into subfamilies. Recently, Beecher (1951a, 1951b, 1953) and Tordoff (1954) have used these features and others which they thought to be of value in an attempt to clarify the relationships of the species involved.

Beecher's work (1951a, 1951b, 1953) on jaw-musculature is a valuable contribution to our knowledge of the anatomy of passerine birds. His myological studies were so thorough and his presentation so detailed that students who disagree with his interpretations can draw their own conclusions. Beecher (1951b:276) points out that there are two basic types of skeletal muscle--those with parallel fibers and those with pinnately arranged fibers. The muscles with pinnate fibers seem to be more efficient, each muscle having a greater functional cross section for its bulk than does one with parallel fibers. He assumes that muscles with parallel fibers are more primitive, phylogenetically, than are those with fibers arranged pinnately. Since his study of the jaw muscles of the Icteridae (1951a) revealed that patterns of jaw-musculature within this Family remain constant regardless of the methods used in procuring food, he assumes that such patterns may be used as indicators of relationship throughout the entire oscinine group. These two assumptions, then, serve as the basis for his hypothesis concerning relationship and phylogeny within this assemblage. Beecher (1951b:278-280; 1953:310-312) maintains that within the Family Thraupidae there are two main lines which lead with almost no disjunction to the Carduelinae and Richmondeninae. The thraupid-richmondenine line involves a shift in the nature of the _m. adductor mandibulae externus superficialis_, which becomes more pinnate in the richmondenines. This results in greater crushing power. The thraupid-cardueline line involves a shift in emphasis from the the _m. adductor mandibulae externus medialis_ to the _m. pseudotemporalis superficialis_ and the forward advance of the insertion of the latter. This, also, promotes greater crushing ability. He states that features of the horny palate and of the plumage provide further evidence of close relationship of these groups. He includes, therefore, the Thraupinae, the Carduelinae, and the Pyrrhuloxiinae (=Richmondeninae) in the Family Thraupidae. Beecher (1953:307) indicates that the patterns of jaw-musculature of the Parulinae (wood warblers) and Emberizinae (buntings) are similar and suggests that the buntings had their origin from the wood warblers. He includes these subfamilies, therefore, in the Family Parulidae.

Beecher's reasoning may be criticized on several points. It may be, as he suggests, that muscles with parallel fibers evolved earlier, phylogenetically, than did muscles with pinnate fibers, but he does not give adequate consideration, it seems to me, to the possibility that parallel fibers may also have evolved secondarily from pinnate fibers. Since Beecher (1951a) found that patterns of jaw-musculature within the Family Icteridae were conservative, he is reluctant to admit the possibility of convergence among any of the other families. Differences in patterns of jaw-musculature are, however, functional adaptations and like the bill, which is also associated with food-getting may be subject to rapid evolutionary change. Finally, in attempting to classify the oscines, he has relied almost entirely on a single character--the pattern of jaw-musculature.

Tordoff's attempts (1954) to clarify the relationships of the fringillids and related species are based chiefly on features of the bony palate. He assumes that since palato-maxillaries seem to be absent in the majority of passerine birds, their occurrence in certain nine-primaried oscine groups indicates relationship among these groups. He points out that these bones, when present, are important areas of origin of the _m. pterygoideus_ which functions in depression of the upper jaw and in elevation of the lower jaw. He assumes, therefore, that palato-maxillaries were evolved to provide for a more effective action of the _m. pterygoideus_. The need for such action could be associated with a seed-eating habit. All richmondenines and emberizines possess palato-maxillary bones either free or fused to the prepalatine bar, but there is no trace of these bones in the carduelines. Carduelines, furthermore, possess prepalatine bars that are characteristically flared anteriorly. This condition does not exist in the richmondenines or in the emberizines.

Tordoff points out, also, that the irregular, erratic migrations of the New World Carduelinae are unlike the more regular migrations of the richmondenines and emberizines. The carduelines, furthermore, are more arboreal in their habits than are these other groups and exhibit a decided lack of nest sanitation during the later stages of nesting, a situation which contrasts with that found in the Richmondeninae and Emberizinae. He suggests, therefore, that the carduelines are not so closely related to the richmondenines and the emberizines as previously has been thought.

Since there are only two cardueline genera, _Loximitris_ and _Hesperiphona_, endemic to the New World and at least 10 genera with many species endemic to the Old World, Tordoff (1954:15) suggests an Old World origin for the carduelines. He strengthens his argument for this hypothesis by pointing out that in features of the bony palate and in habits the carduelines resemble the estrildines of the Family Ploceidae.

Tordoff (1954:29-30) states that the tanagers not only merge with the richmondenines but also grade imperceptibly into the emberizines. He includes, therefore, the Richmondeninae, Emberizinae, and Thraupinae in the Family Fringillidae. He suggests that the carduelines are ploceids, closely related to the Subfamily Estrildinae, on the basis of structure of the bony palate, geographic distribution, social behavior, and habits such as nest-fouling and nest-building.

Tordoff, like Beecher, has based his interpretations chiefly on one feature--structure of the bony palate. Since this feature also is associated with food-getting, the possibilities of convergence of distantly related species with similar habits and divergence of closely related species with different habits may not be excluded.

The hazard of unrecognized adaptive convergence cannot, of course, be excluded from most fields of taxonomic research, but some features of morphology and biochemistry are notably more conservative than others and undergo slower evolutionary change. Such features are often of utmost importance in distinguishing the higher taxonomic categories.

Most ornithologists are aware that, within the Order Passeriformes, patterns of musculature in the leg have evolved at a slow rate and exhibit little variation within the Order. Differences which do occur, therefore, probably are significant, especially those that are consistent between groups of species. As I have pointed out earlier (p. 184), there are no significant differences in leg-musculature between the Richmondeninae, Emberizinae, and Thraupidae. Indeed, it is difficult to define these groups on the basis of leg-musculature. If these groups are of common origin, the lack of distinct boundaries between them is not surprising. A muscular band which extends from the _pars interna_ of the _m. gastrocnemius_ around the front of the knee is present in every emberizine species that I studied and in the Genus _Piranga_. With the exception of _Spiza_ none of the richmondenines possesses this band.

The significant differences in leg-musculature which have been discussed above (pp. 183-184) distinguish the carduelines from the New World finches and tanagers. Even the cardueline _Leucosticte_ and the emberizine _Calcarius_, which resemble one another in general adaptations and in several myological features of the leg (p. 183), agree in significant features of the musculature with the respective groups to which they belong. The carduelines agree in the major features of leg-musculature with the ploceids which I studied.

The use of serological techniques in taxonomic work has two main advantages. The biochemical systems involved in such investigations seem to be relatively slow to change in response to external environmental influences, and the quantitative nature of the results obtained makes possible objective measurement of resemblances among species.

I have pointed out (p. 200) that the carduelines are excluded, serologically, from the distinct assemblage formed by the richmondenines, emberizines, and tanagers. Actually, the carduelines show less serological resemblance to this assemblage than do the estrildines, and most ornithologists agree that the Estrildinae are not at all closely related to the Richmondeninae, Emberizinae, and Thraupidae. _Molothrus_, representing a family (Icteridae) recognized as distinct from the Family Fringillidae, also more closely resembles the fringillid assemblage, serologically, than do the carduelines. Although the Carduelinae constitute a distinct group serologically, they show greater serological resemblance to the estrildines of the Family Ploceidae than to any of the other species tested. At least the carduelines and the estrildines form a group as compact as the subfamilies of the Fringillidae. Thus, the serological data correlate well with those obtained from the study of the leg-musculature.

Present systems of classification include the subfamilies Passerinae and Estrildinae in the Family Ploceidae. _Passer_, however, is less closely related to the estrildines serologically than are the carduelines, and is less closely related to the estrildines than _Molothrus_, an icterid, is to the fringillids. This raises a question as to the homogeneity of the Family Ploceidae as presently recognized by most ornithologists. If the Passerinae and the Estrildinae are placed in a single family, the serological divergence among members of this group is certainly greater than it is in the Family Fringillidae. Additionally, Beecher (1953:303-304) found that the estrildines possess a pattern of jaw-musculature different from those in other ploceids.

The combined evidence from jaw-musculature and serology has caused me to conclude that the estrildines should be excluded from the Family Ploceidae (see below).

In an attempt to clarify the relationships of the Fringillidae and allied groups, I here review briefly the evidence which has been presented. From his studies of jaw-musculature (1951a, 1951b, 1953) Beecher concludes that the Pyrrhuloxinae (=Richmondeninae), the Carduelinae, and the Thraupinae are closely related. He places these groups in the Family Thraupidae. He excludes the Emberizinae from this group and places them with the wood warblers in the Family Parulidae. He suggests that the estrildines constitute a family (Estrildidae) separate from the Family Ploceidae.

From his studies of certain features of the bony palate Tordoff (1954:25-26, 32) concludes that the richmondenines, the emberizines, and the tanagers have a common origin and places these groups in the Family Fringillidae. He excludes the carduelines from this assemblage, suggests that they are closely related to the estrildines, and includes them as the Subfamily Carduelinae in the Family Ploceidae.

In this paper I have presented data obtained from the study of certain features of morphology and biochemistry which I think are less subject to the influence of environmental factors than those features studied by recent workers. It is significant that the data obtained by use of serological techniques and those obtained from the study of leg-musculature point to the same conclusions. On the basis of these data I have drawn several conclusions concerning the relationships of the groups which I studied.

The richmondenines, emberizines, and tanagers are closely related and should be included in a single family, Fringillidae. The Carduelinae and the Estrildinae are closely related subfamilies. Although most recent classifications place the Estrildinae and Passerinae in the Family Ploceidae, the serological evidence indicates that these groups are not closely related. Beecher (1953:303-304) drew the same conclusion from his study of jaw-musculature (see above). I suggest, therefore, that the Carduelinae and the Estrildinae be placed in a family separate from the Ploceidae and that the name Carduelidae (rather than Estrildidae) be used for this group. At present, neither is an accepted family name. Because _Carduelis_ Brisson 1760 is an older name than _Estrilda_ Swainson 1827 and because _Carduelis_ seems to be a centrally located genus in the family, I have chosen the former (although the International Rules of Zoological Nomenclature do not specify that priority must apply in forming family names).

I have been unable to study any of the species included in the subfamilies Fringillinae (not Fringillinae of Tordoff, see 1954:23-24, and below) and Geospizinae of recent classifications; thus these groups have not been discussed above. Beecher (1953:307-308) includes _Fringilla_ in the Subfamily Carduelinae; he includes the geospizines in a separate family, Geospizidae, and states that they are derived from the emberizines. Tordoff (1954:23-24) found that in features of the bony palate _Fringilla_ and the geospizines resemble the emberizines and, on this basis, includes them in the Subfamily Fringillinae.

The Dickcissel, _Spiza americana_, possesses certain features which merit special discussion. Beecher (1951a:431; 1953:309), on the basis of jaw-musculature, considers it an icterid. To be sure _Spiza_ is in many ways an aberrant member of the group to which it is now assigned (Subfamily Richmondeninae). _Spiza_, serologically, is closely related to all species of the richmondenine-emberizine-thraupid assemblage. Within this assemblage its nearest relatives are the richmondenines. _Spiza_ differs from the other richmondenines studied and resembles the emberizines and tanagers in the possession of the muscular band which extends from the _pars interna_ of the _m. gastrocnemius_ around the front of the knee. This band, in _Spiza_, is smaller, however, than in any of the other species. No icterid dissected possesses such a structure. Tordoff (1954:29) states that _Spiza_ is typically richmondenine in palatal structure and makes the suggestion, with which I agree, that _Spiza_ is a richmondenine and may be closely related to the ancestral stock which gave rise to the fringillid assemblage. The serological position of _Spiza_, approximately equidistant from the other fringillids (Figs. 22, 23), and the presence of the small muscular band around the front of the knee constitute evidence supporting the central position of _Spiza_.

After consideration of evidence from the studies of external morphology, ethology, myology, osteology, and serology, I propose here an arrangement of the groups which I have studied and submit for comparison the arrangements (of these groups) proposed by Beecher and Tordoff. The names of subfamilies that I have been unable to study are included in my classification and are placed in brackets.

------------------------+----------------------+----------------------- | Proposed by Tordoff | Proposed by Beecher Here proposed: | (1954) on the basis | (1953) on the basis | of the bony palate: | of jaw-musculature: ========================+======================+======================= FAMILY PLOCEIDAE | FAMILY PLOCEIDAE | FAMILY PLOCEIDAE | | [Subf. Bubalornithinae] |Subf. Bubalornithinae | | | Subfamily Passerinae: |Subfamily Passerinae | Subfamily Passerinae distinguished from the | | Estrildinae by patterns | | of jaw-musculature | | (Beecher, 1953:303-304) | | and on the basis of | | comparative serology of | | saline-soluble proteins.| | | | [Subfamily Ploceinae] |Subfamily Ploceinae | Subfamily Ploceinae | | [Subfamily Viduinae] |Subfamily Viduinae | Subfamily Viduinae | | FAMILY CARDUELIDAE | | | | Subfamily Estrildinae: |Subfamily Estrildinae | FAMILY ESTRILDIDAE similar to the | | Carduelinae in features | | of the bony palate and | | habits (Tordoff, 1954: | | 18-22) and in patterns | | of leg-musculature and | | comparative serology | | of saline-soluble | | proteins. | | | | Subfamily Carduelinae: |Subfamily Carduelinae | [In Thraupidae below] distinguished from the | | Fringillidae by features| | of the palate, | | geographic distribution,| | migration patterns, and | | habits (Tordoff, 1954: | | 14-18) and by patterns | | of leg-musculature and | | comparative serology | | of saline-soluble | | proteins. | | | | FAMILY FRINGILLIDAE: all| FAMILY FRINGILLIDAE | FAMILY PARULIDAE members of this family | | Subfamily Parulinae show similarities in | | Subfamily Emberizinae features of the bony | | palate (Tordoff, 1954: | | 22-23), patterns of | | leg-musculature, and | | in comparative serology | | of saline-soluble | | proteins. | | FAMILY THRAUPIDAE | | Subf. Richmondeninae |Subf. Richmondeninae | Subfamily | | Pyrrhuloxiinae | | Subfamily Thraupinae |Subfamily Thraupinae | Subfamily Thraupinae | | Subfamily Emberizinae |Subfamily Fringillinae| [In Parulidae above] |(including Emberizinae| [Subfamily Fringillinae]| and Geospizinae) | Subfamily Carduelinae | | [Subfamily Geospizinae] | | ------------------------+----------------------+-----------------------

SUMMARY

It has long been recognized that the Family Fringillidae includes some dissimilar groups. Specifically, the relationships of the subfamilies Richmondeninae, Emberizinae, and Carduelinae of the Family Fringillidae are poorly understood. Data from two recent studies, one on patterns of jaw-musculature and the other on features of the bony palate, emphasize the dissimilarity of these subfamilies but have given rise to conflicting concepts of the relationships of subfamilies within the Family.

This paper reports the results of studies involving morphological and biochemical features that I consider less sensitive to external environmental factors than are features which have been studied previously. Patterns of leg-musculature were chosen for study because earlier work showed that muscle patterns in the legs of passerine birds are highly stable and vary but little. Variations, therefore, which are consistent in separating groups of species should be significant. Serological techniques were used because the biochemical systems involved seem to be relatively slow to change in response to environmental influences and because the data obtained may be used in a highly objective manner to measure resemblance among species.

Individual differences in the patterns of leg-musculature were found to be slight and involved mainly the sizes and shapes of muscles. For this reason variations involving origin, insertion, or relative position of a muscle, were judged significant. In leg-musculature the Richmondeninae, the Emberizinae, and the Thraupidae resemble one another closely. Several differences in muscle pattern were found, however, which distinguish these groups from the Carduelinae. The leg-musculature of the carduelines closely resembles that of the Ploceidae.

Serological techniques involved the extraction of saline-soluble proteins from the tissues of the species to be studied. These extracts were carefully processed and were used as antigens. Formolization of the antigens was necessary as a means of preventing denaturation of the proteins by enzymatic activity. Antisera were produced in rabbits. The method of testing involved turbidimetric analysis of the precipitin reaction. Utilizing the values for the precipitin tests a model was constructed which showed the relationships of the eleven species used in these tests. From a study of the model and the data used in its construction, it was determined that the Richmondeninae, Emberizinae, and Thraupidae constitute an assemblage distinct from the other species studied. The Carduelinae are excluded from the assemblage and serologically are most closely related to the Estrildinae. The estrildines, serologically, do not closely resemble _Passer_, Subfamily Passerinae, although recent classifications place these two subfamilies in the Family Ploceidae.

Upon consideration of all evidence now available--from external morphology, ethology, myology, osteology, and serology--several hypotheses regarding the relationships of the groups studied are set forth. The richmondenines, emberizines, and tanagers are closely related subfamilies and are here included in the Family Fringillidae. The Estrildinae and Carduelinae are closely related subfamilies, but neither group is closely related to the Passerinae. The estrildines and carduelines, therefore, are placed in a separate family, the Carduelidae. In some ways, _Spiza_ is an aberrant member of the Subfamily Richmondeninae but should be retained in that subfamily. It is suggested that _Spiza_ is a primitive richmondenine closely related to the ancestral fringillid stock.

LITERATURE CITED

AMERICAN ORNITHOLOGISTS' UNION

1931. Check-list of North American birds. Fourth edition.
Lancaster, Pa., xix + 526 pp.

BEECHER, W. J.

1951a. Adaptations for food-getting in the American blackbirds.
Auk, 68:411-440, 11 figs.

1951b. Convergence in the Coerebidae. Wilson Bull., 63:274-287,
5 figs.

1953. A phylogeny of the oscines. Auk, 70:270-333, 18 figs.

BERGER, A. J.

1952. The comparative functional morphology of the pelvic
appendage in three genera of Cuculidae.
Amer. Mid. Nat., 47:513-605, 29 pls.

BOYDEN, A.

1942. Systematic serology: a critical appreciation.
Physiol. Zool., 15:109-145, 12 figs.

CHAPIN, J. P.

1917. The classification of the weaver-birds. Bull. Amer. Mus.
Nat. Hist., 37:243-280, 10 pls., 9 figs.

CUMLEY, R. W., and IRWIN, M. R.

1941. Pictorial representation of the antigenic differences
between two dove species. Jour. Hered., 32:178-182,
frontispiece, 2 figs.

1941. Interaction of antigens in dove hybrids. Ibid., 429-434,
3 figs.

1944. The correlation between antigenic composition and geographic
range in the Old and New World of some species of _Columba_.
Amer. Nat., 78:238-256, 1 fig.

DEFALCO, R. J.

1942. A serological study of some avian relationships.
Biol. Bull., 83:205-218.

FISHER, H. I.

1946. Adaptations and comparative anatomy of the locomotor
apparatus of New World vultures. Amer. Mid. Nat.,
35:545-727, 13 pls., 28 figs.

GADOW, H., and SELENKA, E.

1891. Voegel, vol. I, Anatomischer Theil. In Bronn's Klassen und
Ordnungen des Thier-Reichs, Sechster Band, Vierte Abtheilung.
Leipzig, 1008 pp., 59 pls.

GARROD, A. H.

1873. On certain muscles in the thigh of birds and their value in
classification. Proc. Zool. Soc. London, Part I:626-644,
6 figs.

1874. On certain muscles in the thigh of birds and their value in
classification. Ibid., Part II:111-123.

GREENBERG, D. M.

1929. The colorimetric determination of serum proteins.
J. Biol. Chem., 82:545-550.

HELLMAYR, C. E.

1935. Catalogue of birds of the Americas. Field Mus. Nat. Hist.,
Zool. ser. 13, pt. 8, vi + 541 pp.

1936. Catalogue of birds of the Americas. Ibid., 13, pt. 9,
v + 458 pp.

1937. Catalogue of birds of the Americas. Ibid., 13, pt. 10,
v + 228 pp.

1938. Catalogue of birds of the Americas. Ibid., 13, pt. 11,
vi + 662 pp.

HOWARD, H.

1929. The avifauna of the Emeryville shellmound. Univ. California
Publ. Zool., 32:301-394, 3 pls., 54 figs.

HUDSON, G. E.

1937. Studies on the muscles of the pelvic appendage in birds.
Amer. Mid. Nat., 18:1-108, 26 pls.

IRWIN, M. R.

1953. Evolutionary patterns of antigenic substances of the blood
corpuscles in Columbidae. Evol., 7:31-50.

IRWIN, M. R., and COLE, L. J.

1936. Immunogenetic studies of species and of species hybrids in
doves, and the separation of species-specific substances in
the backcross. Jour. Exp. Zool., 73:85-108, 1 fig.

LEONE, C. A.

1949. Comparative serology of some brachyuran crustacea and
studies in hemocyanin correspondence. Biol. Bull.,
97:273-286, 3 figs.

1953. Some effects of formalin on the serological activity of
crustacean and mammalian sera. Jour. Immun., 70:386-392,
2 figs.

LIBBY, R. L.

1938. The photronreflectometer--an instrument for the measurement
of turbid systems. Jour. Immun., 34:71-73, 1 fig.

MARTIN, E. P., and LEONE, C. A.

1952. Serological relationships among domestic fowl as shown by
comparisons of protein preparations from corresponding organ
systems. Trans. Kansas Acad. Sci., 55:439-444, 1 fig.

MCGIBBON, W. H.

1945. Further division of contrasting antigens in species hybrids
in ducks. Genetics, 30:252-265.

SASAKI, K.

1928. Serological examination of the blood-relationship between
wild and domestic ducks. Jour. Dept. Agri., Kyushu Imp.
Univ., 2:117-132.

SIMPSON, G. G.