William B. Stallcup

Except for the typographical correction noted below and a few minor changes (missing/extra punctuation) which may have been made but not noted here, the text is the same as presented in the original publication. Some text has been rearranged to restore paragraphs that were split by tables or images. Most of the illustrations have notation to denote the scale compared to the original specimen (example: × 3). Due to the variation in monitor resolution and geometry, the scale is most likely not correct; but is provided as a guide.

Typographical Corrections

Page 187, Table 1 Item 5 : Intavenous => Intravenous

double bar

The relationships of many groups of birds within the Order Passeriformes are poorly understood. Most ornithologists agree that some of the passerine families of current classifications are artificial groups. These artificial groupings are the result of early work which gave chief attention to readily adaptive external structures. The size and shape of the bill, for example, have been over-emphasized in the past as taxonomic characters. It is now recognized that the bill is a highly adaptive structure and that it frequently shows convergence and parallelism.

Since studies of external morphology have failed in some cases to provide a clear understanding of the relationships of passerine birds, it seems appropriate that attention be given to other morphological features, to physiological features, and to life history studies in an attempt to find other clues to relationships at the family and subfamily levels.

This paper reports the results of a study of the relationships of some birds of the Family Fringillidae and is based on the comparative myology of the pelvic appendage and on the comparative serology of saline-soluble proteins. Where necessary for comparative purposes, birds from other families have been included in these investigations.

It has long been recognized that the Fringillidae include dissimilar groups. Recent work by Beecher (1951b, 1953) on the musculature of the jaw and by Tordoff (1954) primarily on the structure of the bony palate has emphasized the artificial nature of the assemblage although these authors disagree regarding major divisions within it (see below).

The Fringillidae have been distinguished from other families of nine-primaried oscines by only one character—a heavy and conical bill (for crushing seeds). Bills of this form have been developed independently in several other, unrelated, groups; as Tordoff (1954:7) has pointed out, Molothrus of the Family Icteridae, Psittorostra of the Family Drepaniidae, and most members of the Family Ploceidae have bills as heavy and conical as those of the fringillids. The ploceids are distinguished from the fringillids by a single external character: a fairly well-developed tenth primary whereas in fringillids the tenth primary is absent or vestigial. Tordoff (1954:20) points out, however, that this distinction is of limited value since in other passerine families the tenth primary may be present in some species of a genus and absent in others. The Genus Vireo is an example. Furthermore, at least one ploceid (Philetairus) has a small, vestigial tenth primary, whereas some fringillids (Emberizoides, for example) possess a tenth primary which is rather large and ventrally placed (Chapin, 1917:253-254). Thus, it is obvious that studies based on other features are necessary in order to attain a better understanding of the relationships of the birds involved.

Sushkin's studies (1924, 1925) of the structure of the bony and horny palates have served as a basis for the division of the Fringillidae into as many as five subfamilies (Hellmayr, 1938:v): Richmondeninae, Geospizinae, Fringillinae, Carduelinae, and Emberizinae.

Beecher (1951b:280) points out that "the richmondenine finches arise so uninterruptedly out of the tanagers that ornithologists have had to draw the dividing line between the two groups arbitrarily." His study of pattern of jaw-musculature substantiates this. He states further that the cardueline finches arise without disjunction from the tanagers. He suggests, therefore, that the two groups of "tanager-finches" be made subfamilies of the Thraupidae and that a third subfamily be maintained for the more typical tanagers. He states that the emberizine finches are of different origin, arising from the wood warblers (1953:307). Beecher (1951a:431; 1953:309) includes the Dickcissel, Spiza americana, in the Family Icteridae, chiefly on the basis of jaw muscle-pattern and the horny palate.

Tordoff (1954:10-11) presents evidence that the occurrence of palato-maxillary bones in nine-primaried birds indicates relationship among the forms possessing them. He points out that all fringillids except the Carduelinae possess palato-maxillaries that are either free or more or less fused to the prepalatine bar. He points out also that in all carduelines, the prepalatine bar is flared at its juncture with the premaxilla, and that the mediopalatine processes are fused across the midline; noncardueline fringillids lack these characteristics. In addition to the above he cites differences between the carduelines and the "other" fringillids in the appendicular skeletons, in geographic distribution, in patterns of migration, and in habits. Tordoff concludes, therefore, that the carduelines are not fringillids but ploceids, their closest affinities being with the ploceid Subfamily Estrildinae. On the basis of palatal structure, the Fringillinae and Geospizinae are combined with the Emberizinae, the name Fringillinae being maintained for the subfamily. The tanagers merge with the Richmondeninae on the one hand and with the Fringillinae on the other. On this basis, Tordoff (1954:32) suggests that the Family Fringillidae be divided into subfamilies as follows: Richmondeninae, Thraupinae, and Fringillinae. The carduelines are placed as the Subfamily Carduelinae in the Family Ploceidae.

From the foregoing, it is apparent that the two most recent lines of research have given rise to conflicting theories regarding relationships within the Family Fringillidae. The purpose of my investigation, therefore, has been to gather information, from other fields, which might clarify the relationships of these birds.

Since the muscle pattern of the leg in the Order Passeriformes is thought to be one of long standing and slow change, any variation which consistently distinguishes one group of species from another could be significant. With the hope that such variation might be found, a study of the comparative myology of the legs was undertaken.

The usefulness of comparative serology as a means of determining relationship has been demonstrated in many investigations. Its use in this instance was undertaken for several reasons: comparative serology has its basis in biochemical systems which seem to evolve slowly; its methods are objective; and its use has, heretofore, resulted in the accumulation of data which seem compatible, in most instances, with data obtained from other sources.

I acknowledge with pleasure the guidance received in this study from Prof. Harrison B. Tordoff of the University of Kansas. I am indebted also to Prof. Charles A. Leone without whose direction and assistance the serological investigations would not have been possible; to Professors E. Raymond Hall and A. Byron Leonard whose suggestions and criticisms have been most helpful in the preparation of this paper; and to T. D. Burleigh of the U. S. Fish and Wildlife Service for gifts of several specimens used in this work. Assistance with certain parts of the study were received from a contract (NR163014) between the Office of Naval Research of the United States Navy and the University of Kansas.

Myological formulae suggested by Garrod (1873, 1874) have been extensively used by taxonomists as aids in characterizing the orders of birds. Relatively few investigations, however, involving the comparative myology of the leg have been undertaken at family and subfamily levels. The works of Fisher (1946), Hudson (1948), and Berger (1952) are notable exceptions.

The terminology for the muscles used in this paper follows that of Hudson (1937), except that I have followed Berger (1952) in Latinizing all names. Homologies are not given since these are reviewed by Hudson. Osteological terms are from Howard (1929).

Specimens were preserved in a solution of one part formalin to eight parts of water. Thorough injection of all tissues was necessary for satisfactory preservation. Most of the down and contour feathers were removed to allow the preservative to reach the skin.

In preparing specimens for study, the legs and pelvic girdle were removed and washed in running water for several hours to remove much of the formalin. They were then transferred to a mixture of 50 per cent alcohol and a small amount of glycerine.

All specimens were dissected with the aid of a low power binocular microscope. Where possible, several specimens of each species were examined for individual differences. Such differences were found to be slight, involving mainly size and shape of the muscles. The size is dependent partly on the age of the bird, muscles from older birds being larger and better developed. The shape of a muscle (whether long and slender or short and thick) is due in part to the position in which the leg was preserved; that is to say, a muscle may be extended in one bird and contracted in another. For these reasons, descriptions and comparisons are based mainly on the origin and insertion of a muscle and on its position in relation to adjoining muscles.

Birds dissected in this study are listed below (in the order of the A. O. U. Check-List):

The descriptions which follow are those of the muscles in the leg of the Red-eyed Towhee, Pipilo erythrophthalmus. Differences between species, where present, are noted for each muscle. The term thigh is used to refer to the proximal segment of the leg; the term crus is used for that segment of the leg immediately distal to the thigh.

Musculus iliotrochantericus posticus (Fig. 2).—The origin of this muscle is fleshy from the entire concave lateral surface of the ilium anterior to the acetabulum. The fibers converge posteriorly, and the muscle inserts by a short, broad tendon on the lateral surface of the femur immediately distal to the trochanter. It is the largest muscle which passes from the ilium to the femur.

Action.—Moves femur forward and rotates it anteriorly.

Comparison.—No significant differences noted among the species studied.

Musculus iliotrochantericus anticus (Fig. 3).—Covered laterally by the m. iliotrochantericus posticus, this slender muscle has a fleshy origin from the anteroventral edge of the ilium between the origins of the m. sartorius anteriorly and the m. iliotrochantericus medius posteriorly. The m. iliotrochantericus anticus is directed caudoventrally and inserts by a broad, flat tendon on the anterolateral surface of the femur between the heads of the m. femorotibialis externus and m. femorotibialis medius and just distal to the insertion of the m. iliotrochantericus medius.

Action.—Moves femur forward and rotates it anteriorly.

Comparison.—No significant differences noted among the species studied.

Musculus iliotrochantericus medius (Fig. 3).—Smallest of the three iliotrochantericus muscles, this bandlike muscle has a fleshy origin from the ventral edge of the ilium just posterior to the origin of the m. iliotrochantericus anticus. The fibers are directed caudoventrally, and the insertion is tendinous on the anterolateral surface of the femur between the insertion of the other two iliotrochantericus muscles.

Action.—Moves femur forward and rotates it anteriorly.

Comparison.—No significant differences noted among the species studied.

Musculus iliacus (Figs. 4, 5).—Arising from a fleshy origin on the ventral edge of the ilium just posterior to the origin of the m. iliotrochantericus medius, this small slender muscle passes posteroventrally to its fleshy insertion on the posteromedial surface of the femur just proximal to the origin of the m. femorotibialis internus.

Action.—Moves femur forward and rotates it posteriorly.

Comparison.—No significant differences among the species studied.

Musculus sartorius (Figs. 1, 4).—A long, straplike muscle, the sartorius forms the anterior edge of the thigh. The origin is fleshy, half from the anterior edge of the ilium and from the median dorsal ridge of this bone and half from the posterior one or two free dorsal vertebrae. The insertion is fleshy along a narrow line on the anteromedial edge of the head of the tibia and on the medial region of the patellar tendon.

Action.—Moves thigh forward and upward and extends shank.

Comparison.—In Loxia and Spinus, only one-third of the origin is from the last free dorsal vertebra. In Hesperiphona, Carpodacus, Pinicola, and Leucosticte, only one-fifth of the origin is from this vertebra.

Musculus iliotibialis (Fig. 1).—Broad and triangular, this muscle covers most of the deeper muscles of the lateral aspect of the thigh. The middle region is fused with the underlying femorotibialis muscles. In the distal half of this muscle there are three distinct parts; the anterior and posterior edges are fleshy and the central part is aponeurotic. The origin is from a narrow line along the iliac crests—from the origin of the m. sartorius, anteriorly, to the origin of the m. semitendinosus posteriorly. The origin is aponeurotic in the preacetabular region but fleshy in the postacetabular region. The distal part of the muscle is aponeurotic and joins with the femorotibialis muscles in the formation of the patellar tendon. This tendon incloses the patella and inserts on a line along the proximal edges of the cnemial crests of the tibiotarsus.

Action.—Extends crus.

Comparison.—In Vireo the central aponeurotic portion of this muscle is absent.

Musculus femorotibialis externus (Fig. 2).—Covering the lateral and anterolateral surfaces of the femur, this large muscle has a fleshy origin from the lateral edge of the proximal three-fourths of the femur. The origin separates the insertion of the m. iliotrochantericus anticus from that of the m. ischiofemoralis and, in turn, is separated from the origin of the m. femorotibialis medius by the insertions of the m. iliotrochantericus anticus and m. iliotrochantericus medius. Approximately midway of the length of the femur this muscle fuses anteromesially with the m. femorotibialis medius. Distally, the m. femorotibialis externus contributes to the formation of the patellar tendon which inserts on a line along the proximal edges of the cnemial crests of the tibiotarsus.

Action.—Extends crus.

Comparison.—No significant differences noted among the species studied.

Musculus femorotibialis medius (Figs. 2, 4).—The origin of this muscle, which lies along the anterior edge of the femur, is fleshy from the entire length of the femur proximal to the level of attachment of the proximal arm of the biceps loop. Laterally this muscle is completely fused for most of its length with the m. femorotibialis externus and contributes to the formation of the patellar tendon, which inserts on a line along the proximal edges of the cnemial crests of the tibiotarsus. Many of the fibers, nevertheless, insert on the proximal edge of the patella.

Action.—Extends crus.

Comparison.—No significant differences noted among the species studied.

Musculus femorotibialis internus (Fig. 4).—One of the most superficial muscles lying on the medial surface of the thigh, this muscle is divided, especially near the distal end, into two parts, lateral and medial. The origin of the lateral part is fleshy from a line on the medial surface of the femur; the origin begins proximally at a point near the insertion of the m. iliacus. The medial, bulkier part of the muscle has a fleshy origin on the medial surface of the lower one-third of the femur. The two parts fuse to some extent above the points of insertion and insert on the medial edge of the head of the tibia.

Action.—Rotates tibia anteriorly.

Comparison.—Two parts of this muscle variously fused; otherwise, no significant differences in the species studied.

Musculus piriformis (Fig. 3).—This muscle is represented by the pars caudifemoralis only, the pars iliofemoralis being absent in passerine birds as far as is known. The pars caudifemoralis is flat, somewhat spindle-shaped, and passes anteroventrally from the pygostyle to the femur. The origin is tendinous from the anteroventral edge of the pygostyle, and the insertion is semitendinous on the posterolateral surface of the shaft of the femur about one-fourth its length from the proximal end.

Action.—Moves femur posteriorly and rotates it in this direction; moves tail laterally and depresses it.

Comparison.—No significant differences noted among the species studied.

Musculus semitendinosus (Figs. 2, 3, 5).—The origin from the extreme posterior edge of the posterior iliac crest of the ilium is fleshy and is aponeurotic from the last vertebra of the synsacrum and the transverse processes of several caudal vertebrae. The straplike belly passes along the posterolateral margin of the thigh. Immediately posterior to the knee, the muscle is divided transversely by a ligament. That portion passing anteriorly from the ligament is the m. accessorius semitendinosi (here considered a part of the m. semitendinosus) and is discussed below. The ligament continues distally in two parts; one part inserts on the medial surface of the pars media of the m. gastrocnemius and the other part fuses with the tendon of insertion of the m. semimembranosus.

The m. accessorius semitendinosi extends anteriorly from the above mentioned ligament to a fleshy insertion on the posterolateral surface of the femur immediately proximal to the condyles.

Action.—Moves femur posteriorly, flexes the crus and aids in extending the tarsometatarsus.

Comparison.—No significant differences noted among the species studied.

Musculus semimembranosus (Figs. 3, 4, 5).—This straplike muscle passes along the posteromedial surface of the thigh. The origin is semitendinous along a line on the ischium, from a point dorsal to the middle of the ischiopubic fenestra to the posterior end of the ischium, and from a small area of the abdominal musculature posterior to the ischium. The insertion is by means of a broad, thin tendon on a ridge on the medial surface of the tibia immediately distal to the head of this bone. The tendon of insertion passes between the head of the pars media and pars interna of the m. gastrocnemius and is fused with the tendon of the m. semitendinosus.

Action.—Flexes crus.

Comparison.—No significant differences noted among the species studied.

Musculus biceps femoris (Fig. 2).—Long, thin, and somewhat triangular, this muscle lies on the lateral side of the thigh just underneath the m. iliotibialis. Its origin is from a line along the anterior and posterior iliac crests underneath the origin of the m. iliotibialis. Anterior to the acetabulum the origin is aponeurotic, and the edge of this aponeurosis passes over the proximal end of the femur. The origin posterior to the acetabulum is fleshy. The most anterior point of origin is difficult to ascertain but it lies near the center of the anterior iliac crest. The most posterior point of origin is immediately dorsal to the posterior end of the ilioischiatic fenestra. Behind the knee the fibers of this muscle converge to form the strong tendon of insertion which passes through the biceps loop, under the tendon of origin of the m. flexor perforatus digiti II, and inserts on a small tubercle on the posterolateral edge of the fibula at the point of the tibia-fibula fusion.

The biceps loop is tendinous and the distal end attaches to a protuberance on the posterolateral edge of the femur at the proximal edge of the external condyle. The proximal end attaches to the anterolateral edge of the femur immediately proximal to the distal end of the loop, which extends posterior to the femur. The distal arm of this loop is connected with the tendon of origin of the m. flexor perforatus digiti II by a strong tendon.

Action.—Flexes crus.

Comparison.—No significant differences noted among the species studied.

Musculus ischiofemoralis (Fig. 3).—Short and thick, this muscle arises directly from the lateral surface of the ischium between the posterior iliac crest and the ischiopubic fenestra. The area of origin extends to the posterior edge of the ischium. The insertion is tendinous on the lateral surface of the trochanter opposite the insertion of the m. iliotrochantericus medius.

Action.—Moves femur posteriorly and rotates it in this direction.

Comparison.—No significant differences noted among the species studied.

Musculus obturator internus (Figs. 4, 7).—Lying on the inside of the pelvis and covering the medial surface of the ischiopubic fenestra, is this flat, pinnate, leaf-shaped muscle. The origin is fleshy and is from the ischium and pubis around the edges of this fenestra; none of the fibers arises from the membrane stretched across the fenestra. Anteriorly the fibers converge and form a strong tendon that passes through the obturator foramen and inserts on the posterolateral surface of the trochanter of the femur.

Action.—Rotates femur posteriorly.

Comparison.—No significant differences noted among the species studied.

Musculus obturator externus (Fig. 7).—Short and fleshy, this muscle consists of two parts which are not easily separable but which may be traced throughout its length. The parts are more nearly distinct at the origin. The dorsal part arises directly from the ischium along the dorsal edge of the obturator foramen. The larger ventral part arises directly from the anterior and ventral edges of the obturator foramen. The fibers of the dorsal part pass anteriorly, cover the tendon of the m. obturator internus laterally, and insert on the trochanter around the point of insertion of the latter muscle. The fibers of the ventral part pass parallel with the tendon of the m. obturator internus and insert on the trochanter immediately distal and posterior to the tendon of the latter muscle.

Action.—Rotates femur posteriorly.

Comparison.—In Passer, Estrilda, Poephila, Hesperiphona, Carpodacus, Pinicola, Leucosticte, Spinus and Loxia, this muscle is undivided and, in its position, origin, and insertion, resembles the ventral part of the bipartite muscle described above. The origin is from the anterior and ventral edges of the obturator foramen and the insertion is on the trochanter of the femur immediately distal and posterior to the insertion of the m. obturator internus. In all other genera examined, the muscle is bipartite. In Chlorura the dorsal part is larger and better developed than it is in the other genera.

Musculus adductor longus et brevis (Figs. 3, 4, 5).—Consisting of two distinct, straplike parts, this large muscle lies on the medial surface of the thigh, posterior to the femur.

The pars anticus has a semitendinous origin on a line that extends posteriorly from the posteroventral edge of the obturator foramen to a point half way across the membrane that covers the ischiopubic fenestra. The insertion is fleshy along the posterior surface of the femur from the level of the insertion of the m. piriformis distally to the medial surface of the internal condyle.

The pars posticus originates by a broad, flat tendon on a line across the posterior half of the membrane that covers the ischiopubic fenestra. The insertion is at the point of origin of the pars media of the m. gastrocnemius on the posteromedial surface of the proximal end of the internal condyle of the femur. There is a broad tendinous connection with the proximal end of the pars media of the m. gastrocnemius. The anterior edge of the pars posticus is overlapped medially by the posterior edge of the pars anticus.

Action.—Flexes thigh; may flex crus also and may extend tarsometatarsus.

Comparison.—In Vireo olivaceous, the origin of this muscle does not extend the length of the ischiopubic fenestra. The origin, furthermore, is along the dorsal edge of the ischiopubic fenestra and not from the membrane covering the fenestra. Finally, in this species, the origin of the pars posticus is fleshy.

Musculus tibialis anticus (Figs. 2, 5).—Lying along the anterior edge of the crus, a part of this muscle is covered by the m. peroneus longus. The origin is by two distinct heads, each of which is pinnate. The anterior head arises directly from the edges of the outer and inner cnemial crests. The posterior head arises by a short, strong tendon from a small pit on the anterodistal edge of the external condyle of the femur. This tendon and the proximal end of the muscle pass between the head of the fibula and the outer cnemial crest. The two heads of the muscle fuse at a place slightly more than one-half of the distance down the crus. At the distal end of the crus this muscle gives rise to a strong tendon which passes under a fibrous loop immediately proximal to the external condyle in company with the m. extensor digitorum longus and which passes between the condyles of the tibia and inserts on a tubercle on the anteromedial edge of the proximal end of the tarsometatarsus.

Action.—Flexes tarsometatarsus.

Comparison.—No significant differences noted among the species studied.

Musculus extensor digitorum longus (Figs. 3, 5, 8).—Slender and pinnate, this muscle lies along the anteromedial surface of the tibia. The origin is fleshy from most of the region between the cnemial crests and from a line along the anterior surface of the proximal fourth of the tibia. Approximately two-thirds of the distance down the crus the muscle gives rise to the tendon of insertion which passes through the fibrous loop near the distal end of the tibia in company with the m. tibialis anticus. The tendon then passes along beneath the supratendinal bridge at the distal end of the tibia, traverses the anterior intercondylar fossa, and passes beneath a bony bridge on the anteromedial surface of the proximal end of the tarsometatarsus. The tendon continues along the anterior surface of the tarsometatarsus to a point immediately above the bases of the toes and there gives rise to three branches, one to the anterior surface of each foretoe. The insertions of each branch are on the anterior surfaces of the phalanges as shown in Fig. 8.

Action.—Extends foretoes.

Comparison.—This muscle is weakly developed in Leucosticte and Calvarius; the belly is slender and extends only half way down the crus before giving rise to the tendon of insertion. The functional significance of this variation is difficult to understand. The convergence in muscle pattern shown by these two genera, however, is in all probability the result of similarities in behavior patterns. These birds perch less frequently than do the other birds studied. Thus, the toes are neither flexed nor extended as often; the smaller size of the m. extensor digitorum longus may have resulted in part from this lessened activity. Except for the variations just noted, there are no significant differences among the species studied; even the rather complex patterns of insertion are identical.

Musculus peroneus longus (Fig. 1).—Relatively thin and straplike, this muscle lies on the anterolateral surface of the crus and is intimately attached to the underlying muscles. The part of the origin from the proximal edges of the inner and outer cnemial crests is semitendinous but the part of the origin from the lateral edge of the shaft of the fibula is tendinous. Approximately two-thirds the distance down the crus the muscle gives rise to the tendon of insertion. Immediately above the external condyle of the tibiotarsus this tendon divides. The posterior branch inserts on the proximal end of the lateral edge of the tibial cartilage. The anterior branch passes over the lateral surface of the external condyle to the posterior surface of the tarsometatarsus and there unites with the tendon of the m. flexor perforatus digiti III.

Action.—Extends tarsometatarsus and flexes third digit.

Comparison.—No significant differences noted among the species studied.

Musculus peroneus brevis (Figs. 2, 3).—Lying along the anterolateral surface of the tibia, this slender, pinnate muscle arises from a fleshy origin along this surface and along the anterior surface of the fibula from a point immediately proximal to the insertion of the m. biceps femoris to a point approximately two-thirds of the way down the crus. Near the distal end of the tibia the muscle gives rise to the tendon of insertion that passes through a groove on the anterolateral edge of the tibia just above the external condyle. Here the tendon is held in place by a broad fibrous loop and passes under the anterior branch of the tendon of insertion of the m. peroneus longus and inserts on a prominence on the lateral edge of the proximal end of the tarsometatarsus.

Action.—Extends tarsometatarsus and may abduct it slightly.

Comparison.—No significant differences noted among the species studied.

Musculus gastrocnemius (Figs. 1, 4).—The largest muscle of the pelvic appendage, it covers superficially all of the posterior surface, most of the medial surface, and half of the lateral surface of the crus. The muscle originates by three distinct heads.

The pars externa covers the posterolateral surface of the crus, is intermediate in size between the other two heads, and arises by a short, strong tendon from a small bony protuberance on the posterolateral side of the distal end of the femur immediately proximal to the fibular condyle. The tendon is intimately connected with the distal arm of the loop for the m. biceps femoris.

The pars media is the smallest of the three heads and lies on the medial surface of the crus. The head of the pars media is separated from the pars interna by the tendon of insertion of the m. semimembranosus and originates by a short, strong tendon from the posteromedial surface of the proximal end of the internal condyle of the femur. The proximal portion of the pars media has tendinous connections with the tendon of the m. semitendinosus and with the pars posticus of the m. adductor longus et brevis.

The pars interna is the largest of the three heads and covers most of the medial surface of the crus. This head in its proximal portion is distinctly divided into anterior and posterior parts, the former overlapping the latter medially. The origin of the posterior part is fleshy from the anterior half of the tibial head. Some of the fibers of the anterior part arise directly from the inner cnemial crest while its remaining fibers arise from the patellar tendon (Fig. 1) and form a band that extends around the anterior surface of the knee, covering the insertion of the m. sartorius.

Approximately half way down the crus, the three heads give rise to the tendon of insertion, the tendo achillis, which passes over and is tightly bound to the posterior surface of the tibial cartilage. The insertion is tendinous on the posterior surface of the hypotarsus and along the posterolateral ridge of the tarsometatarsus. This tendon seems to be continuous with a fascia which forms a sheath around the posterior surface of the tarsometatarsus holding the other tendons of this region firmly in the posterior sulcus.

Action.—Extends tarsometatarsus.

Comparison.—Study of the pars externa and pars media reveals no significant differences among the species dissected. The pars interna, however, is subject to some variation which is described below.

The two parts of the m. gastrocnemius are most distinct in Vireo. Icterus, Molothrus, Richmondena, Guiraca, and Passerina lack the fibrous band that passes around the front of the knee. In Spiza this band of fibers is smaller than in the other species.

In Leucosticte, although the pars interna is undivided, there is a band of fibers which extends around the front of the knee (see discussion, p. 183).

Musculus plantaris (Fig. 5).—Small and slender, this muscle lies on the posteromedial surface of the crus, beneath the pars interna of the m. gastrocnemius and originates by fleshy fibers from the posteromedial surface of the proximal end of the tibia immediately distal to the internal articular surface. The belly extends approximately one-sixth of the way down the crus and gives rise to a long, slender tendon that inserts on the proximomedial edge of the tibial cartilage.

Action.—Extends tarsometatarsus.

Comparison.—No significant differences noted among the species studied.

Musculus flexor perforatus digiti II (Figs. 3, 9).—This is a slender muscle which lies on the lateral side of the crus beneath the pars externa of the m. gastrocnemius and is intimately connected anteromedially with the m. flexor digitorum longus and posteromedially with the m. flexor hallucis longus. The origin is by a strong tendon from the lateral surface of the external condyle of the femur at the point of origin of the m. flexor perforans et perforatus digiti II. This tendon serves also as the origin of the anterior head of the m. flexor hallucis longus. The tendon connects also by a broad tendinous band with the distal arm of the loop for the m. biceps femoris and by a similar band with the lateral edge of the fibula immediately distal to the head. The tendon of insertion passes distally, perforates the tibial cartilage near its lateral edge, traverses the middle medial canal of the hypotarsus (Fig. 6), and passes distally to the foot. At the distal end of the tarsometatarsus the tendon is held against the medial surface of the first metatarsal by a straplike sheath. The tendon then passes over a sesamoid bone between the first metatarsal and the base of the second digit and is bound to this bone by a sheath. The tendon inserts mainly along the posteromedial edge of the proximal end of the first phalanx of the second digit, although the termination is sheathlike and covers the entire posterior surface of this phalanx. This sheathlike termination is perforated by the tendons of the m. flexor perforans et perforatus digiti II and the branch of the m. flexor digitorum longus that inserts on the second digit.

Action.—Flexes second digit.

Comparison.—In Vireo this muscle is larger and more deeply situated than it is in the other species examined and has no connection with the m. flexor hallucis longus.

Musculus flexor perforatus digiti III (Fig. 5).—Long and flattened, this muscle lies on the posteromedial side of the crus beneath the m. gastrocnemius. The belly is tightly fused laterally with the belly of the m. flexor hallucis longus and posteriorly with the belly of the m. flexor perforatus digiti IV. The origin is by a long, strong tendon from a small tubercle just medial to, and at the proximal end of, the external condyle of the femur. Below the middle of the crus this muscle terminates in a strong tendon which perforates the tibial cartilage near its lateral edge. In this region the tendon is sheathlike and wrapped around the tendon of the m. flexor perforatus digiti IV. These two tendons together pass through the posterolateral canal of the hypotarsus (Fig. 6). Immediately distal to the hypotarsus the two tendons separate, and the tendon of the m. flexor perforatus digiti III receives a branch of the tendon of the m. peroneus longus. The tendon passes distally over the surface of the second trochlea, and its insertion is sheathlike on the posterior surface of the first phalanx, and on the proximal end of the second. In the area of insertion this tendon is perforated by that of the m. flexor perforans et perforatus digiti III and by that of the m. flexor digitorum longus to the third digit.

Action.—Flexes digit III.

Comparison.—In Passer, Estrilda, Poephila, Hesperiphona, Carpodacus, Pinicola, Leucosticte, Spinus, and Loxia the edges of the sheathlike tendon are thickened at the points of insertion, so that the tendon appears to have two branches which insert along the posterolateral edges of the first phalanx and are connected medially by a fascia.

Musculus flexor perforatus digiti IV (Fig. 3).—Extending along the posterior edge of the crus, this slender muscle lies beneath the m. gastrocnemius. The belly is fused with those of the m. flexor hallucis longus and m. flexor perforatus digiti III. Its origin is fleshy from the intercondyloid region of the distal end of the femur and has a few fibers arising from the tendon of origin of the m. flexor perforatus digiti III. Near the distal end of the crus the muscle gives rise to the strong tendon of insertion which perforates the tibial cartilage near its lateral edge and in this region is ensheathed by the tendon of the m. flexor perforatus digiti III. The two tendons pass together through the posterolateral canal of the hypotarsus (Fig. 6). The tendon continues distally along the tarsometatarsus and the posterior surface of digit IV. The tendon bifurcates at approximately the middle of the first phalanx. A short lateral branch inserts on the posterolateral edge of the proximal end of the second phalanx. The long medial branch is perforated by a branch of the m. flexor digitorum longus; the distal end is flattened, has thickened edges, and inserts over the posterior surfaces of the distal end of the second phalanx, and over the proximal end of the third phalanx.

Action.—Flexes digit IV.

Comparison.—No significant differences noted among the species studied.

Musculus flexor perforans et perforatus digiti II (Figs. 2, 9).—Small and spindle-shaped, this muscle lies on the posterolateral side of the crus immediately beneath the pars externa of the m. gastrocnemius. The origin is fleshy and arises in company with the m. flexor perforans et perforatus digiti III from a point on the posterolateral surface of the distal end of the femur between the point of origin of the pars externa of the m. gastrocnemius and the fibular condyle. The belly extends approximately one-fourth of the way down the crus and gives rise to the tendon of insertion which passes distally and superficially through the posterior edge of the tibial cartilage. The tendon traverses the posteromedial canal of the hypotarsus (Fig. 6) and continues along the posterior surface of the tarsometatarsus. Between the first metatarsal and the base of the second digit the tendon is enclosed by the medial surface of a sesamoid bone. This tendon then perforates that of the m. flexor perforatus digiti II at the level of the first phalanx and in turn is perforated by the tendon of the m. flexor digitorum longus at the proximal end of the second phalanx. The insertion is on the posterior surface of the second phalanx.

Action.—Flexes digit II.

Comparison.—In Passer, Estrilda, Poephila, Hesperiphona, Carpodacus, Pinicola, Leucosticte, Spinus, and Loxia the proximal portion of this muscle is more intimately connected with the posterior edge of the m. flexor perforans et perforatus digiti III than it is in the other species examined.

Musculus flexor perforans et perforatus digiti III (Fig. 2).—Long and pinnate, this muscle lies on the lateral surface of the crus beneath the m. peroneus longus and pars externa of the m. gastrocnemius. There are two distinct heads. The origin of the anterior head is fleshy from the proximal edge of the outer cnemial crest and from the internal edge of the distal end of the patellar tendon. The posterior head arises by a tendon from the femur in company with the m. flexor perforans et perforatus digiti II, is connected also with the tendon of origin of the m. flexor perforatus digiti II, and is loosely attached to the head of the fibula. Fibers from the belly of the muscle attach throughout its length to the lateral edge of the fibula, and the muscle is tightly fused also with adjacent muscles. The tendon of insertion is formed approximately one-half the way down the crus. The tendon perforates the posterior surface of the tibial cartilage and passes through the posteromedial canal of the hypotarsus (Fig. 6). At the base of the third digit the tendon ensheathes that of the m. flexor digitorum longus and the two together perforate the tendon of the m. flexor perforatus digiti III. Immediately distal to this perforation the tendon of the m. flexor perforans et perforatus digiti III ceases to ensheath that of the m. flexor digitorum longus. The latter passes beneath that of the former. Near the distal end of the second phalanx the tendon of the m. flexor digitorum longus perforates that of the m. flexor perforans et perforatus digiti III. The latter inserts on the posterior surface of the distal end of the second phalanx and the proximal end of the third.

Action.—Flexes digit III.

Comparison.—In Passer, Estrilda, and Poephila, and in all the cardueline finches examined the proximal portion of this muscle is more intimately connected with the anterior edge of the m. flexor perforans et perforatus digiti II than it is in the other species examined.

Musculus flexor digitorum longus (Figs. 3, 5).—This strong, pinnate muscle is deeply situated along the posterior surfaces of the tibia and fibula. There are two distinct heads of origin. The lateral head arises by means of fleshy fibers from the posterior edge of the head of the fibula. The medial head arises by means of fleshy fibers from the region under the ledgelike external and internal articular surfaces of the proximal end of the tibia. Neither head has any connection with the femur in contrast to the condition, described by Hudson (1937: 46-47) in the crow, Corvus brachyrhynchos, and in the raven, Corvus corax. Near the point of insertion of the m. biceps femoris the two heads fuse. The common belly is attached by fleshy fibers to the posterior surface of the tibia and fibula for two-thirds of the distance down the crus. Near the distal end of the crus the muscle terminates in a strong tendon which passes deeply through the tibial cartilage and traverses the anteromedial canal of the hypotarsus (Fig. 6). About midway down the tarsometatarsus this tendon becomes ossified. Immediately above the bases of the toes it gives rise to three branches, one to the posterior surface of each of the foretoes. These branches perforate the other flexor muscles of the toes as described in the accounts of those muscles and insert as follows: The branch to digit II inserts on the base of the ungual phalanx and by a stout, tendinous slip on the distal end of the second phalanx (Fig. 9). The branch to digit III inserts on the base of the distal end of the third phalanx and a stronger slip to the distal end of the second or proximal end of the third. The branch to digit IV inserts on the base of the ungual phalanx, with one tendinous slip to the distal end of the third phalanx and another to the distal end of the fourth.

Action.—Flexes foretoes.

Comparison.—No significant differences noted among the species studied.

Musculus flexor hallucis longus (Fig. 3).—Situated immediately posterior to the m. flexor digitorum longus, the belly of this large, pinnate muscle is intimately connected anteriorly to that of the m. flexor perforatus digiti II. The m. flexor hallucis longus arises by two heads which are separated by the tendon of insertion of the m. biceps femoris. The smaller anterior head arises from the same tendon as does the m. flexor perforatus digiti II. The larger posterior head arises by means of fleshy fibers from the intercondyloid region of the posterior surface of the femur along with the m. flexor perforatus digiti III and IV. The two heads join just distal to the point of insertion of the m. biceps femoris. There is no trace of a tendinous band connecting the two heads as there is in the crow and in the raven (Hudson, 1937:49). Near the distal end of the shank the muscle gives rise to a strong tendon which perforates the tibial cartilage along its lateral edge and passes through the anterolateral canal of the hypotarsus (Fig. 6). The tendon crosses over to the medial surface of the tarsometatarsus, passes distally, and perforates the sheathlike tendon of the m. flexor hallucis brevis between the first metatarsal and the trochlea for digit II. The tendon continues along the posterior surface of the hallux and has a double insertion; the main tendon attaches to the base of the ungual phalanx and a smaller branch inserts on the distal end of the proximal phalanx.

Action.—Flexes hallux.

Comparison.—In Vireo this muscle has only the posterior head of origin and is not connected with the m. flexor perforatus digiti II. The muscle is proportionately smaller and weaker than in any of the other species studied.

Musculus extensor hallucis longus (Fig. 4).—One of the smallest muscles of the leg, the origin is fleshy from the anteromedial edge of the proximal end of the tarsometatarsus. The belly is long and slender and terminates distally in a slender tendon which passes distally along the posterior surfaces of the first metatarsal and the first digit. The insertion is on the base of the ungual phalanx. Near the distal end of the proximal phalanx, the tendon passes between two thick bands of fibro-elastic tissue which insert also on the ungual phalanx. These bands of tissue function as automatic extensors of the claw.

Action.—Extends hallux; action must be slight.

Comparison.—In Vireo this muscle is proportionately larger and better developed than it is in any of the other species examined.

Musculus flexor hallucis brevis (Fig. 4).—This minute muscle has a fleshy origin from the medial surface of the hypotarsus. The short belly terminates in a weak, slender tendon which passes down the posteromedial surface of the tarsometatarsus and into the space between the first metatarsal and the trochlea for digit II. In this region the tendon envelops the tendon of the m. flexor hallucis longus and inserts on the distal end of the first metatarsal and on the proximal end of the first phalanx of the first digit.

Action.—Flexes hallux; action must be slight.

Comparison.—The small size of this muscle makes it exceedingly difficult to study. The muscle is larger in Vireo than in any of the other species examined. This may be correlated with the smaller size of the m. flexor hallucis longus in this species. The muscle does not seem to be so well developed in the cardueline finches as it is in the other species.

Musculus abductor digiti IV (Fig. 2).—Extremely small, delicate and difficult to demonstrate, this muscle arises in a fleshy origin immediately from underneath the posterior edge of the external cotyla of the tarsometatarsus. The tendon of insertion is long and slender and inserts along the lateral edge of the first phalanx of digit IV.

Action.—Abducts digit IV.

Comparison.—No significant differences noted among the species studied.

Musculus lumbricalis.—Semitendinous throughout its length, this muscle arises from the ossified tendon of the m. flexor digitorum longus at a point immediately proximal to the branching of this tendon. The insertion is on the joint pulleys and capsules at the base of the third and fourth digits.

Action.—Hudson (1937:57) states that: "Meckel (vide Gadow—1891, p. 204) considered this muscle as serving to draw the joint pulley behind in order to protect it from pinching during the bending of the toes. It perhaps also tends to flex the third and fourth digits."

Comparison.—No significant differences noted among the species studied.

Simpson (1944:12) and others have emphasized that different parts of organisms evolve at different rates. Beecher (1951b:275) in stating that "... the hind limb is very similar in muscle pattern throughout the Order Passeriformes and seems to have become relatively static after attaining a high level of general efficiency ..." implies that the muscle pattern of the leg must be one of long standing and slow change. This concept was emphasized by Hudson (1937) who found but little variation in muscle pattern among members of the several families of passerine birds. The concept is further confirmed by the present investigation. The intricate patterns of origin and of insertion seem to remain almost the same throughout the order in spite of adaptive radiation which has occurred.

Two major differences in patterns of leg-musculature, however, were found among the species studied, and these differences are significant since they are consistent between subfamilies. The muscles involved are the m. obturator externus and the pars interna of the m. gastrocnemius.

The m. obturator externus is bipartite, consisting of dorsal and ventral parts, in the passerine species studied by Hudson (1937) and in all of the species examined by me except the ploceids and the cardueline finches. In the ploceids and cardueline finches this muscle is undivided and resembles in its position, origin, and insertion only the ventral portion of the muscle found in the other birds studied. It is difficult to imagine what advantage or disadvantage might be associated with the bipartite or with the undivided condition. The action of this muscle is to rotate the femur (right femur clockwise, left femur counterclockwise), and certainly the greater mass of the bipartite muscle could lend greater strength to such action. The possible significance of this is discussed below.

Abd. dig. IV M. abductor digiti IV
Acc. M. accessorius semitendinosi
Add. long. M. adductor longus et brevis
Anterolat. can. Anterolateral canal of hypotarsus
Anteromed. can. Anteromedial canal of hypotarsus
Bic. fem. M. biceps femoris
Bic. loop Loop for m. biceps femoris
Ext. cot. External cotyla
Ext. dig. l. M. extensor digitorum longus
Ext. hal. l. M. extensor hallucis longus
Fem. tib. ext. M. femorotibialis externus
Fem. tib. int. M. femorotibialis internus
Fem. tib. med. M. femorotibialis medius
F. dig. l. M. flexor digitorum longus
F. hal. brev. M. flexor hallucis brevis
F. hal. l. M. flexor hallucis longus
F. p. et p. d. II M. flexor perforans et perforatus digiti II
F. p. et p. d. III M. flexor perforans et perforatus digiti III
F. per. d. II M. flexor perforatus digiti II
F. per. d. III M. flexor perforatus digiti III
F. per. d. IV M. flexor perforatus digiti IV
Gas. M. gastrocnemius
Iliacus M. iliacus
Il. tib. M. iliotibialis
Il. troc. ant. M. iliotrochantericus anticus
Il. troc. med. M. iliotrochantericus medius
Il. troc. post. M. iliotrochantericus posticus
Int. cot. Internal cotyla
Isch. fem. M. ischiofemoralis
Midmed. can. Midmedial canal of hypotarsus
Obt. ext. M. obturator externus
Obt. int. M. obturator internus
P. ant. Pars anticus
P. ext. Pars externa
P. int. Pars interna
P. med. Pars media
P. post. Pars posticus
Per. brev. M. peroneus brevis
Per. long. M. peroneus longus
Pirif. M. piriformis
Plan. M. plantaris
Posterolat. can. Posterolateral canal of hypotarsus
Posteromed. can. Posteromedial canal of hypotarsus
Sar. M. sartorius
Semim. M. semimembranosus
Semit. M. semitendinosus
Tib. ant. M. tibialis anticus
Tib. cart. Tibial cartilage

Fig. 6. Pipilo erythrophthalmus. Proximal end of left tarsometatarsus and the hypotarsus, × 4.

Fig. 7. Pipilo erythrophthalmus. Lateral view of proximal end of left femur and a portion of the pelvis, × 3.5.

Fig. 8. Pipilo erythrophthalmus. Upper surfaces of the phalanges of the foretoes of the left foot showing insertions of the M. extensor digitorum longus, × 3.

Fig. 9. Pipilo erythrophthalmus. Medial view of the second digit of the left foot, showing insertions of the flexor muscles, × 3.

The division of the pars interna of the m. gastrocnemius into anterior and posterior parts has not been reported by previous authors yet the division is quite distinct in those birds in which it occurs. Hudson (1937:36) points out that in some non-passerine birds the pars interna is double, but that in these species the m. semimembranosus inserts between the two parts. This is not the condition in those species studied by me. Only the ploceids and the cardueline finches in the present investigation fail to show such a division. The undivided muscle in these birds resembles, in its origin and position, the posterior portion of the muscle found in those species showing the bipartite condition. The greater mass of the bipartite muscle probably makes possible a stronger extension of the tarsometatarsus.

Thus, the divided or undivided conditions of the m. obturator externus and the pars interna of the m. gastrocnemius seem to be correlated with the degrees of strength of certain movements of the leg. It is conceivable that these differences in structure are correlated with the manner in which food is obtained, the birds having the bipartite muscles being those which spend the most time on the ground searching and scratching for seeds and other sorts of food. Yet, in Leucosticte, a cardueline, and in Calcarius, an emberizine, whose foraging habits are rather similar, the structure is unlike. Leucosticte does resemble the emberizines and also Piranga and Spzia in the extension of a band of muscle fibers from the pars interna of the m. gastrocnemius around the front of the knee. A band of muscle fibers of this sort strengthens the knee joint and gives still more strength to the pars interna. This condition has been reported in a number of birds by Hudson (1937) and is, in all probability, an adaptation for greater strength of certain leg movements. The development of this band in Leucosticte seems to parallel that in the other birds studied and does not indicate relationship, since in Leucosticte this band arises from the undivided muscle which (as stated above) resembles only the posterior portion of the bipartite muscle described for the other birds. In the latter, the muscular band arises from the anterior part of the muscle.

Minor differences in muscle pattern, like those already mentioned, are consistent also between subfamilies, but correlation of these minor differences with function is difficult. There is the implication, however, that in all the groups except the carduelines and ploceids, the emphasis is on greater strength and mobility of the leg. In the carduelines that were studied the origin of the m. sartorius does not extend so far craniad as in the other species. In the latter, at least half of the origin is from the last one or two free dorsal vertebrae; in the carduelines no more than one third of the origin is anterior to the ilium. It is conceivable that the more craniad the origin, the stronger the forward movement of the thigh would be.

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.

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.

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°C., the specimen was placed in a Waring blender with 0.9 per cent aqueous solution of NaCl buffered with M/150 K₂HPO₄ and M/150 Na₂HPO₄ 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°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°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°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

[A] Antiserum prepared against formolized antigen.