SEA-URCHINS. 485 



of the scar left by the apical system indicates a considerable b;ick- 

 Wiird prolongation of tluit structure. I have before nie a series 

 of ten specimens from the Pea Grit of Crickley Hill (near Chel- 

 tenham), which are presiuiiably the young of Pyyaster semisulcatus ; 

 but I am unable to find any sa,tisfactory distinctions between 

 them and the type of //. bo/iei. If there is any appreciable 

 ditlei'ence, it consists in the fact that the perij)roct does not 

 project so far into the posterior interambulacrum in the Pyyasters 

 as does the " scar of the apical disc " in the Hemipedina. It 

 seems hardly possible that, so early in the history of both orders, 

 heterogenetic homoeomorpiiy could have reached such a degiee of 

 perfection, and I am thei-efore strongly of the opinion that 

 " llemi'pedina " honei is a Fijyaster, and almost certainly a young 

 form of /*. semisulcatus. 



It thus seems established that Pi/gaster is intimately related to 

 .some primitive, probably Liassic, Diademoid. It is impracticable, 

 in the present state of our knowledge, to search for the actual 

 (jeneric ancestor ; but if the choice were to lie between J)imh- 

 inopsis and Hemijiediaa, 'the former would seem to possess the 

 stronger claim to recognition. As defined by Lambert (52), 

 Diademopsis is distingnihhed from Ilemipediiia by the presence 

 of pronounced secondary tubercles in the interambulacra. 

 Bather (59) has shown that the distinction is not so absolute as 

 Lambert's diagnosis would suggest, but the fact remains that, 

 among the earlier species of the genei'a, there is a more strongly 

 developed tendency to a mnltituberculate character in JJia- 

 demopsis. As P)/<juster is also a multitul)erculate form, the 

 alliance with I)iademoj)sis would seem natural, but I do not feel 

 justified in expressing a positive opinion on the matter, lieyond 

 the statement that the immediate ancestor of the Holectypoida 

 inust surely have been a ],)iademoid. 



3. The Pygasteridse and Connlida^. 



The three subgenera of Pygaster sens. lat. mark three stages 

 in the evolution of that genus. Pygaster sens. str. is undoubtedly 

 the most primitive type. Megapygus shows an advance in two 

 directions. The periproct is undergoing a change of shape pre- 

 liminary to its actual separation fi'om the apical system, and the 

 tuberculation is a.ssuming slight irregularity of arrangement. 

 IJoth these features point towards " Irregnlai'ity." Macropyg^is, 

 Avhich appeared at abont the same horizon as Megapygus, shows a 

 similar character in its periproct, but the tubei'Cidation, instead 

 of becoming superficial and irregular, shows a deepening of the 

 scrobicules, and a corresponding reduction of the miliary surface. 

 The distinction from the Meyapygus vmhreUa group is not veiy 

 gi-eat in appearance, but seems impoitant in its results. I regard 

 the two subgenera as parallel lines springing from the common 

 ancestor Pygaster sens. str. 



Pileus is undoubtedly a short-lived offshoot from the Pygaster- 



