100 THE VOYAGE OF H.M.S. CHALLENGER. 



tabulate, and styles in both forms of them. The dactylozooids of Archistylaster were 

 devoid of the knobbed tentacles, these were, however, retained by the gastrozooid. 

 The oonanoia were included in hollows in the ccenosteum. 



In Sporadopora, the most ancestral Stylasterid at present known, the styles of the 

 dactylopores have disappeared, and they only reappear apparently by reversion in 

 AUopora and Stylaster. Rudimentary tabulae are present in Sporadopora and Plio- 

 boihrus, but disappear in succeeding genera. In Pliobothrus the margins of the dac- 

 tylopore mouths are raised up and prolonged into small tubuli, and the genus would 

 thus lead to Errina, where the tubuli become nariform, were it not that in Pliobothrus 

 the style of the gastrozooid is lost, and that the gastrozooid is devoid of tentacles and 

 flask-shaped : a condition occurring again only in the most highly specialised members 

 of the family Astylus and Oryptohelia. Distichopora appears to have been derived 

 directly from some form allied to Errina. 



Two separate modifications of the nariform projections of Errina are presented by 

 Porella and Spinipora, in both which genera further complication ensues hj the 

 differentiation of two kinds of dactylozooids. 



The process of the formation of cyclo-systems is seen in all stages in different parts 

 of the surface of the single species AUopora subviolacea, as will be seen by reference 

 to Saville Kent's figures, 1 or to the diagrams given on Plate I. of the present Memoir, 

 figs. 10, 11, and 12. In this coral five or six dactylopores are grouped in a circle 

 around a single centrally-placed gastropore. In some groups all the pores are simply 

 circular (fig. 10). In others, shallow grooves, often only just indicated, lead radially 

 from the dactylopores towards the gastropore. In others, these grooves are well 

 marked and deep, and a complete cyclo-system is formed. It appears probable that 

 this condition has been brought about by the continual bending inwards of the dactylo- 

 pores to convey food to the gastropore. The grooves have been the result of the 

 attempts of these zooids to reach the gastrozooid when further and further retracted. 

 Thus, in most Alloporas and all Stylasters, all the pores have come to form regular 

 cyclo-systems, in which the mouths of the dactylopores are drawn out into elongate 

 chambers, and their tubular prolongations reduced to mere rudiments in many cases. 

 At a very short distance below the surface in AUopora subviolacea the pores are found 

 to be in all the systems still entirely independent (fig. 12), and this condition is 

 maintained at greater depths in all Stylasters. It is to be noted that in becoming 

 so remarkably modified into elongate slit-like cavities, the dactylopores of Stylasteridae 

 with cyclo-systems follow an ancesti*al tendency to modification, for these elongate pores 

 are, taken separately, closely similar in form to the nariform dactylopores of Errina 

 and its allies : they only have the lips of all the projections directed radially out- 

 wards. 



1 Proc. Zool. Soc, 1871, pi. xxv. fig. 2 a. 



