vestigial). Posteriorly the pouches decreased in size, while 

 the more anterior pairs were somewhat smaller in a gradi- 

 ent starting with the third or fourth pouch. The oral cavity 

 extended down to the ventral wall of the carapace all 

 around. It was perforated posteriorly by a large opening for 

 the dorsal aorta, below which was a canal for the esophagus 

 and below this, one for the ventral aorta. The esophageal 

 and ventral aorta openings were sometimes confluent. There 

 was no opening for a nasohypophyseal connection with the 

 oral cavity. 



There is no evidence of ventral endocranial elements nor 

 of a tongue apparatus. 



In many of the cephalaspids, there was a lateral spine 

 extending back to either side of the trunk. Just inside of this 

 spine was a lobe-like fin. The endoskeleton was exposed in 

 the area of the base of this fin and is called the pectoral 

 sinus. The walls of this sinus are perforated by vascular and 

 nerve foramina. 



The osteostracan type was peculiar in having mandibular, 

 hyoid (spiracular), and eight pairs of branchial pouches, 

 whereas in the lamprey, only the glossopharyngeal (first 

 branchial) and vagus innervated pouches are present. In the 

 hagfish, only vagus innervated pouches are present. A 

 premandibular pouch has been postulated for the fossil 

 forms but this does not agree with the nerve distribution. 

 The pouches of the osteostracan were peculiar in that they 

 extended further forward in the head, the more anterior 

 ones lying well in front of the eye. The mouth was not 

 suctorial. 



Anaspid The anaspids are known only from faint outlines 

 and impressions but appear to have been free-swimming 

 forms with possibly a weakly suctorial mouth. The head of 

 Birkenia (Figure 5-29 B) was covered by many small scales 

 which were probably imbricated, not overlapped. The eye 

 was surrounded by larger plates, and there was a perforated 

 pineal plate in the roof between the eyes. Anterior to the 

 pineal plate was a single nasal opening. Well behind and 

 below the lateral eyes were eight gill openings. The mouth 

 is not indicated in any of the fossils but is assumed to have 

 been a vertical slit or oval. The round area in the cheek be- 

 hind the mouth was perhaps related to a suctorial function. 

 No endoskeleton is known, although a tooth plate believed 

 to belong to the tongue apparatus has been described for 

 Pharyngolepis. The amount of dermal ossification varies from 

 the scaled Birkenia to the nearly naked Lasamus. The arrange- 

 ment of plates and scales does not suggest any basic pattern, 

 and the number of gill openings varies from 6 to 15. 



Heferostracan The earliest known vertebrates, those of 

 the Middle Ordovician, are believed to have been heter- 

 ostracans on the basis of the microstructure of their armor 

 fragments (see scales. Chapter 8). The actual outline of these 

 fishes is known from Upper Silurian and Lower Devonian 

 species. 



The head-trunk armor consisted of several distinct plates 

 ( Figure 5-32). The cornual plate may have extended as a 

 pectoral fin. The mouth structure is poorly known but is 

 described for a few species. The bilateral nasal openings prob- 

 ably lay just inside the mouth margin, and the hypophys- 

 eal pouch is assumed to have been in the roof of the mouth. 



The endoskeleton is not known, but molds of the inner 

 surface of the carapace show the positions of the gill pouches, 

 the pineal organ, and the two semicircular canals of the 

 inner ear. In its relation to the inner ear, the most anterior 

 pouch was the hyoid, or spiracular. The pouches, varying in 

 number from 7 to 14, all opened to the outside through a 

 single port, well back on the side of the carapace. The endo- 

 cranium was apparently cartilaginous and basically like 

 that of a cephalaspid. 



The drepanaspid (Figure 5-33) had the same pattern of 

 plates as the typical pteraspid, but the dorsal and ventral 

 plates were surrounded by an area of small polygonal units. 

 The carapace was not sharply delimited from the tail. 



In other heterostracans, such as the poraspid, the dorsal 

 and ventral plates were fused to form single elements above 

 and below; the eye no longer opened through an orbital 

 plate but rather through a notch in the margin of the dorsal 

 shield. The branchial plate remained separate or was fused 

 to the dorsal shield. The earliest heterostracan, the astras- 

 pid, has the entire carapace made up of a mosaic of small 

 polygonal plates. In the drepanaspid there is evidence of the 

 fusion of these to form the large plates of the pteraspid. 



Stensio has compared the head structure of the pteraspid 

 with that of the myxinid, but the fossil material does not 

 support some of his assumptions, such as those regarding the 

 nasohypophyseal duct and the occurrence of tentacles on 

 the mouth margin The two semicircular canals of the inner 

 ear disagree with the condition in the hagfish. Thus, the 

 position of the myxinid is not clarified by these comparisons. 



Generol observations The armored condition of many 

 of the fossil agnaths has been viewed as peculiar in the light 

 of the naked living forms. The single-piece carapace of some 

 fossils also presents questions as to growth. It is assumed 

 that the younger stages were naked or had small denticles 

 in the skin and that the armor was acquired with full growth. 



Among the heterostracans, the poraspids acquired their 

 armor when fully grown, but the pteraspids developed 

 their plates at ever earlier points in their life history, and 

 the individual plates increased peripherally in size. The 

 earliest heterostracan, the astraspid, was covered by a mosaic 

 of small plates and thus had no restrictions as to growth. 

 Similarly, the anaspid appears to have developed its dermal 

 scales and plates early in its life history, and the increase 

 in size of these dermal elements was accomplished by mar- 

 ginal growth. 



One can conclude that the most primitive agnaths were 

 naked or were covered with a shagreen of small denticles 

 (like the shark, Squalus) and that plates or a solid carapace 



AGNATH FISHES • 137 



