450 H. L. Hawkins — Morphology of Echinoidea. 



evidence, the preponderance of probability favours an Holectypoid 

 origin for Pyrina ; so that, provisionally at least, Pyrina and 

 Echinoneus must be included among the Holectypoida. 



In the second paper of this series an account was given of the 

 details of ornament developed in Conulus, particularly in 

 C. albogalerus. The following notes deal with the main tubercles 

 only, and treat of those morphogenetically, so that in a sense they 

 supply particulars that were omitted in the previous account. But 

 the purpose of this return to the subject of the ornament of Conulus 

 is different. As Jackson has clearly shown, the " law of localized 

 stages of development " is often very well illustrated by the tubercu- 

 lation of the interambulacra of Echinoids. The plates towards the 

 peristome, being ontogenetically young, show recapitulatory stages 

 in ornament which (subject, of course, to modifications introduced 

 after their formation) give valuable aid in tracing the ancestry of the 

 form concerned. The plates of the mid-zone show the full "species- 

 characters", while those near the apex, being morphogenetically 

 young, provide a less convincing, but recognizable and partly 

 prophetic, type of recapitulation. With a view to discovering the 

 probable relationships of Conulus I have analysed interambulacral 

 columns of three types (probably distinct species) from the English 

 Chalk (see Figure, p. 445). A similar analysis of Pyrina desmoulinsi is 

 given to show the similarity, and difference, between the genera in 

 this matter. The species of Conulus are arranged in stratigraphical 

 order — C. sp. (a large, subglobular, form) is from the zone of 

 R. cuvieri near Reigate, C. subrotundus (a tall variety) from the zone 

 of T. lata near Wallingford, and C. albogalerus (a large, acutely 

 conical form with a wide periproct — Gravesend type) from the 

 sub-zone of C. albogalerus, about 20 feet below the zone of 

 Uintacrinus, at Whitway, Hants. In all types supernumerary 

 tubercles occur near the ambitus. These supernumeraries, which 

 occur equally in Discoidea cylindrica, are clearly hypertrophied 

 secondaries, and have no place in the fundamental scheme of tuber- 

 culatum They are interesting as showing one method whereby 

 tuberculation may become multiplied and disordered, while the 

 apicad pattern in C. sp. (and to a less degree in the two other species) 

 show a second method producing a comparable result. 



The interambulacral plates figured have been selected from the 

 same columns and parts of columns as far as possible. For convenience 

 in spacing on the page the columns of each species run horizontally. 

 In the first vertical row the proximal orad plates are shown. Next 

 comes a plate from the " mid-zone " of the adoral surface, and in the 

 middle row a plate from the ambitus. The two right-hand columns 

 represent plates from the "mid-zone" of the adapical surface, and the 

 apicad plates. It has proved more satisfactory to select the plates in 

 this way than to choose mid-zonal plates in the literal sense. The 

 ambitus in all these forms has various specializations associated with 

 it, and the disposition of the tubercles is usually very different on the 

 two surfaces of the test. It is true that this tendency somewhat 

 vitiates the application of the law of localized stages, but it cannot 

 be avoided in forms with sharp ambital angles. In Pyrina, where 



