THE SUDDEN ORIGIN OF NEW TYPES 409 



proceed beyond the basipterygial stage. ^ The limb-girdles, 

 which show great variations in form, are clearly adaptive 

 structures, and were apparently segmented off from the proxi- 

 mal radials or basipterygium ; and although limb-girdles are 

 present in the more specialised, but still very primitive Pleura- 

 canthus, the two halves of the limb-girdle remain distinct. 

 In proportion to the exigencies of an increased demand for 

 support and leverage for the fins, the limb-girdles increased 

 in size by dorsal and ventral out-growths, and subsequent 

 union in the median plane. It is, however, only in Elasmo- 

 branchs and Dipnoi, not in Crossopterygians, that actual 

 union and fusion of the two halves takes place. Not until 

 this stage had been reached, giving a suitable amount of support, 

 could fins or paddles become adapted, by further specialisation 

 and reduction, to the optimum necessary for crawling on the 

 sea-shore. 



It may be remarked parenthetically that in Pleiir acanthus the 

 anal fins, fig. 6 (which may have some use for crawling on the 

 sea-bottom), show, if anything, a more superficial resemblance- — 

 by a variable fusion of radials — to the pentadactylate limb than 

 the pectoral and pelvic limbs do ; and since this is the case in 

 a fish which is so generalised a type that it might " with very 

 little modification become either a Selachian Dipnoan or Crosso- 

 pterygian,"^ it furnishes us with a possible clue as to the analo- 

 gous way in which the pentadactylate limbs could become 

 differentiated from the paired fins. A marked reduction is also 

 to be seen in the dorsal fin of the Crossopterygian Glypiolepis 

 (almost handlike), fig. 7, in which the proximal radials are 

 united at their base into a single piece — the basipterygium — but 

 are free distally.^ 



In Elasmobranchs and Crossopterygians^ the basal elements 



' R. C. Osburn, "The Origin of Vertebrate Limbs," .<4««. New York Acad. 

 Set. 1906-7, xvii. No. 2, Pt. ii. p. 419, has shown that in Cestracion the rays 

 and basals begin to appear before the girdle, and this is supported by the 

 pateontological evidence of Cladoselache. But all three structures are differen- 

 tiated out of the same band or layer of mesenchyme. 



* A. Fritsch, Fauna dcr Gaskohle, 1883-1901. 



' A. Smith- Woodward, Vertebrate Pala-ontology, p. 32, 1S98. 



* A. Smith-Woodward, British Museum Catalogue 0/ Fossil Fishes, ii. p. 335, 

 1891. 



* In the Crossopterygian Eusthenopteron Foordi, Wht. Fig. 8, of the Upper 

 Devonian of Canada (Whiteaves, Trans. Roy. Soc. Canada, vi. 1888, p. 77 ; and 



27 



