PORIFERA. II. 



correct, would be polyaxial spicules, seems to me to be of no importance, as the question might very 

 well be of a secondary splitting of the axis of the original monactinal spicule. Otherwise I shall 

 remark that it may be proved that the tooth of the chela can be traced back to a bent axial part, 

 while the teeth of the aucorse, in spite of the fact that their form quite corresponds to that of the 

 tooth of the chela, are, perhaps, not axial bendings, but are formations which have, all of them, also 

 the middle tooth, arisen in another way, through growth by apposition. I cannot, however, at present 

 establish this as a sure fact. If it should prove correct, Levinsen of course is not right in designating 

 the teeth of aucorse as axial bendings, but the principal difference between chela; and ancorae will 

 then be still greater. In this connection it may be well to remind of the fact that ancorae may be 

 found provided with an even number of teeth, without any middle tooth. 



Besides the division into chelae and ancorse Levinsen, in the paper quoted, subdivides chela; 

 into two forms, chelae palmata; and arcuatse. The former have a straight, or most frequently only 

 slightly curved shaft, and rather large alae forming together an almost triangular or oval plate, often 

 with a deep notch below; the tooth is often rather broad. In the other form, chelae arcuatse, the shaft 

 is most frequently rather strongly curved, the alse are highly indented below and only attached to 

 the shaft for a comparatively short way, thereby getting a somewhat tooth-like appearance; the tooth 

 is most frequently rather narrow. To be sure, these two forms are not principally different, but differ 

 only in form, and forms may also be met with that may only with difficulty be referred to either of 

 them; but generally they form two well-marked groups, and are also characteristic for certain genera. 

 Therefore there is every reason to keep this division. 



By all the authors who acknowledge only one form, chelae, these are generally described as 

 «tridentate=> or .palmate . This division is a complete mistake; the forms called « palmate > are gener- 

 allv chelae palmata;, but in the term «tridentate> are comprised both chelae arcuatse and ancorse. 

 Thus Ridley and Dendy and Topsent use generally these terms; nay, in Topseut's work on the 

 sponges of the Belgic Antarctic expedition we even meet again the old term bidente». Under Des- 

 macidon setifer it is said in the text of the chelse «dentes ou palmes >, and on PI. Ill, fig. 6 a series is 

 figured of which it is said in the explanation of the figures, «a, b, formes rudimentaires» ; this is cor- 

 rect, the figures show two developmental forms; then the continuation runs, *c, isochele palme, d, iso- 

 chele dente, c', d', formes intermediates >. All the figures belong to one form, a typical palmate chela, 

 c is seen from the front, d from the side, while c' and d' are two a little contort specimens, and one 

 end is therefore seen a little from the front, the other a little from the side. 



Besides the division made by Levinsen of the forms belonging here, I further divide ancorae 

 into two classes which I call ancorse spatuliferae and unguiferae. In the first form the shaft is most 

 frequently straight or slightly curved, and the teeth are comparatively large and broad; their number 

 is very frequently three, but there may be more. In ancorae unguiferae the shaft is most frequently 

 more curved, the teeth are comparatively small and pointed, their number is three to nine, generally 

 more than three. These two forms of ancorse are only different in form, and transitional forms occur, 

 but the two forms are most frequently well-marked and characteristic of particular genera. 



In the palmate chelae the ends may be either like or unlike each other, and according to this 

 they are either isochelse or anisochelse; in the arcuate chelae, the ends, as far as known, are always 



