562 DR. J. F. GEMMILL ON THE 



4. The terminal, the first amijulacral, and the mouth-angle 

 ossicles sliow certain very interesting features in earlier or later 

 development. 



Terminals. — Usually the first indication of the terminal in each 

 i-ay is a pair of small calcifications, to which a third (median) is 

 frequently added, while others appear later at the sides. The 

 adult terminal is formed as in Crihrella (p. 560) by fusion of 

 a number (probably usually five) of these early calcifications. 

 This has taken place by the time the young starfish is 3 mm. in 

 diameter. 



Amhulacral Skeleton. — In each i-ay on each side of the middle 

 line, a skeletal plate develops (a) proximal to the first sucker-foot, 

 (h) between the first and second mucker-feet, and (c) between the 

 second and third sucker-feet, and so on in order. In my paper 

 on the development of Solaster eadeca I followed Ludwig (6) in 

 numbering these as the first, second, and third ambulacrals 

 (A , A^, A.,), and also in reckoning the mouth-angle plates as 

 being the first of the adambulacrals (Adj, Ludwig). The A^ plates 

 give rise to what are usually called in the adult the proximal 

 processes of the first ambulacral vertebrje. Careful examination, 

 confirmed by slow maceration experiments, shows that both in 

 Solaster endeca and in Grossaster fafj^osus the mouth-angle 

 plates (Adj, Ludwig) are usually continuous in their calcification 

 with the corresponding Aj plates. The joint which separates 

 them in the adult must therefore be of secondary foi'mation. In 

 view of the circumstance just noted, and of the further facts 

 (1 ) that palfeozoic Asterozoa have no ambulacra,! bars proximal to 

 the first sucker-feet (Spencer, 12, p. 30), (2) that in many recent 

 starfish the two ambulacral bars in this position in each ray are 

 somewhat widely separated, (3) that each A^ plate wlien it first 

 appears is nearer a mid-interradial than a mid-radial line, w^e 

 may raise the question whether the proximal processes of the first 

 ambulacral vertebrae are not phylogenetically buttress-extensions 

 from the mouth-angle plates towards the middle line of the ray. 

 This view assumes that ontogenetically the formation of the 

 buttress-extension is now somewhat antedated as compared with 

 that of the mouth-angle plate. Such " heterochronicity " is by 

 no means uncommon in other life-histories. 



lY. Stichaster roseus. 



The Stichasteridge and the Asteriidte are included by Sladen 

 among the cryptozonates (11, p. xxxvi). They agree in the 

 general details of their internal anatomy as well as (1) in having 

 the gills not confined to the purely abactinal surface, (2) in 

 having quadriserially arranged sucker-feet, (3) in having the 

 actinosfcomial mai'gin formed by the ambulacial plates, (4) in 

 having pedicellariee both of the forcipiform and of the forficiform 

 type. They diflfer, however, in that among the Stichasteridte the 

 marginals are fairly definite and touch one anothei-, while the 

 plates of the abactinal skeleton are tesselate, and arranged more 



