200 



Bashford Dean Memorial Volume 



Head foof 



very near the median line, immediately in front of the ventral shield (Text-figure 87). 

 As Broili pointed out, very probably these arches are a part of the visceral skeleton. 



Together with the majority of scientists, the present writer is of the opinion that the 



gill apparatus in the Arthrodira was placed under the head shield and that the split be- 



tween PSO and AL is the gill opening. It is hardly possible that the body carapace served 



,-----.^^ as a protector for the branchial region (cor- 



/' "v responding to the opercular region in other 



fishes). In forms like Acanthaspida from 

 Spitsbergen (Heintz 1929.1, .2) the body 

 carapace is very long, making it unlikely that 

 the gill openings would have been placed be- 

 hind the body carapace, since they would 

 have been pushed unnaturally far backward. 

 The same would occur in Coccosteus angus' 

 tus, described by Bryant (1929). On the con- 

 trary, in forms like Heterostius the whole body 

 carapace is shortened and thickened, and there 

 is no room for'gills behind it (Heintz 1930.2). 



It has been remarked that there is too 

 little room for the gill apparatus underneath 

 the head shield. If we compare the heads of 

 the Arthrodira with the heads of other fossil 

 and recent fishes, it is easily seen that in the 

 majority of forms the place for the branchial 

 apparatus is not larger than that in the Ar- 

 throdira. In fact, behind the posterior part of SO and PSO and the front part of AL and 

 IL, there is sufficient room to place an effective branchial apparatus. 



Another circumstance makes the location of the gills in this region very probable. 

 The surface of IL, as mentioned before, is the only place in the whole of Dinichthys, 

 showing a distinctly superficial ornamentation. This ornamentation has been previously 

 described (See Plate VIII, figure 22). We must suppose that it is a preserved skin 

 part which covered this part of the bone. If we remember that in many other Arthrodires 

 (Traquair 1890.4, Jaekel 1902, Heint? 1929.1, .2) this part of IL is sculptured like all 

 other carapace plates — that is, it must be placed superficially — this supposition becomes a 

 certainty. In that case there must have been an opening along IL, otherwise we could not 

 find skin developed here. This opening can only be considered as a branchial opening. 

 The remains of the visceral skeleton found by Broili are also placed in the same region. 

 All these circumstances make it still more probable that the branchial apparatus in the 

 Arthrodira was placed in the hind lower corner of the head. 



The other question is: How was this apparatus formed? We have no remains to 

 help us in solving this problem. We only know that the branchial apparatus was rela- 



Text-figure 87- 



Fragments of the head roof, the ventral shield 



and part of the body of Lunas^is heroldi. 



Drawn to show bd, body; gl, gill arches (?), 

 (After Broili, 1929). 



