the site of attachment of the intercostal or thoracic artery, and these values may come to be used as an index in specific levels of evolution....

"The medial migration of the thoracic artery appears to have some phylogenetic significance as yet not understood."

The six types of thoracic arteries described in Glenny's classification were distinguished as follows (Glenny, 1955:544):

Possibly the thoracic artery has undergone migration but apparent differences in its origin might well be due to differences in other vessels of the thoracic area. Additionally, there seems to be no reason to assume that the lateral position of the thoracic artery is the primitive one, or that the medial is the derived position, as is implied by the phrase "medial migration." Although the lateral site of attachment (type 1) is predominant in the lower orders of birds, and the medial attachment is found primarily in Passeriformes, a fact which may indicate that type 1 is the more primitive, it must nevertheless be kept in mind that a sequence of a single morphological character does not necessarily represent the phylogenetic sequence of the character itself (see Mayr, 1955:41).

Also, a given arterial arrangement might be independently derived more than once. If such has been the case, similarities in arterial arrangements in different taxa would sometimes be "chance similarities," that is to say, "resemblance in characteristics developed in separate taxa by independent causes and without causal relationship involving the similarity as such" (Simpson, 1961:79).

The particular arrangement of the arteries of the thoracic area also seems to be of limited value as a clue to taxonomic relationships. If the origin of any artery is determined by skeletal and muscular features, as I suggest, the artery perhaps ought not be considered as a separate character, but as part of a "character complex" that varies as a unit (see Mayr, Linsley, and Usinger, 1953:123). The skeleton offers a potential fossil record for consideration. Changes in the skeleton and muscles, great enough to affect the blood vessels, would probably be detected more easily than would the resulting vascular changes. Also, I did not find as much individual variation in the skeleton and muscles in the area studied as I did in the vascular system. In other words, within the bounds established by the skeletal and muscular features, the artery still exhibited individual variation in exact origin.

SUMMARY

The origin, distribution, and individual variation of the thoracic and coracoid arteries, and their branches, have been studied in four species of the family Columbidae (pigeons) and in five species of the family Hirundinidae (swallows). These arteries are described for Scardafella inca (Inca Dove) and Progne subis (Purple Martin). Muscles that are supplied by these vessels, and muscles the particular configuration of which seems to effect the arrangement of the arteries have also been described. Correlation of the arteries observed with those named and described by other workers has been attempted.

In most of the vessels studied there is a high degree of individual variation, but few interspecific differences were noticed within either family. Differences in the arteries of the thorax between the two families are described by discussing the resulting different origins of the thoracic artery. In swallows the thoracic artery arises from either the subclavian artery or the coracoid artery, whereas in pigeons it arises from the pectoral trunk. This difference in site of attachment seems to be a result of differences between the two families in muscular and skeletal elements of the thorax.

The particular site of attachment of the thoracic artery is of limited value as a taxonomic character. Several considerations influenced this conclusion. (1) If the location of the artery is determined by skeletal and muscular elements, these associated structures must be considered taxonomically as a "character complex" (a set of characters varying as a unit). (2) Even within the bounds established by the skeleton and muscles, the artery displays a high degree of individual variation in exact origin. (3) A given arterial arrangement could have been derived independently many times. (4) Because differences are defined relative to other likewise variable vessels, supposed similarities or differences in the one artery may be artifacts of the system of description.

My findings and interpretations do not support previous suggestions that the thoracic artery has undergone a mediad migration, and that the various sites of attachment of that vessel may come to represent various levels of evolution. The primitive site of attachment of the vessel is unknown, and it seems to me that it has not been sufficiently demonstrated that the vessel has undergone any "migration."