12 



Memhraniponi Xornidiu n. sp. (PI. XXII, fif^s. 5 a — r> c) is also a partial i-xccplion 

 to the above rule, as the ohliqiiely ascending frontal i)art of the terminal wall 

 can he s|)lil into two, which on the other hand does not seem to he the case 

 with the horizontal pari of this wall. 



Of the tw-o other walls we may first consider the frontal (or oral), which 

 presents the most nnmerous modifications and is therefore systematically the most 

 im])()rtant. In a preliminary paper' I have proposed to divide the cheilostom- 

 alous linjozod into four groups: Mithicoxlecja, Acanlhoslcgit, (]oU\xlv(jii and Cjiumro- 

 steya, which are to a larger or smaller degree based on the structure of this 

 wall. Of these the first three correspond with the division instituted in this work 

 under the name Anasca, which covers all the chcilostomatous linjozoa with no 

 com|)ensation sac, while the fourth corresponds to the Asi-ojtlioni piovided with 

 such a sac. Though I only intend to keej) two of these names for systematic 

 divisions, it would be practical to use adjectives corresponding to all the four 

 names, in order through them to indicate essential dill'erences in the structure of 

 the frontal wall. This may namely be rej)resenfed only by a membrane (mala- 

 costegous Cheilostomat(i) in a largei- or smaller ])art ol' its extent, or is (|uite cal- 

 cified (stereostegous Ch.). In the latter case the chalk cover may be arched (ca- 

 marostegous Ch), or it may be depressed and encircled by projecting margins (coilo- 

 stegous) Ch.). Finally, above the covering membrane there may be a chalk co\er 

 consisting of two rows of hollow spines connected with each other in dill'ercnl 

 ways (acanthostegous ^.Vi.), and lastly, we may just recall tliat the calcareous skel- 

 eton itself may either be a Gymnocyst, a Cryi)locyst or a compound of both. 



Before we try to give a view over the ap|)earance and extension of the Gym- 

 nocyst and the Cryptocyst within the dilferent families of the cheilostomatous 

 Krijozoa, we may discuss some criteria, which might lielj) to determine the ])res- 

 ence of these two kinds of skeleton in cases where there is no possibility of 

 deciding the question directly, namely, by observation of the membrane which 

 should always cover the Cryptocyst. This ap])lies not only in most cases to the dried 

 liryozod, but the thin membrane is also torn away from many colonies pre- 

 served in spirit, and I may mention as an instance, that I have had to examine 

 many spirit specimens of Iisclt(iioi(hs .hich'soni before finding the covering mem- 

 brane. The lateral and oral spines always, as mentioned before, spring from the 

 border of a Gynmocyst, and therefore we can with certainty take it for granted 

 that every calcification which appears within such a spine-bearing border or 

 from a corresponding border in a non-spinous s])ecies is a (Cryptocyst. Ilarmer'- 



' 56, p. 2; - lU, p. ;i2G. 



