12 



2. Sterrospira. (PI. IV, fig. 9). — r, u — y.v.\ fig. 14; fig. 15a — <?. PI. V, fig. ia — 3; fig. 7a — S; 

 fig. 11). As stated before, there is a great difference in aspect between the early stages and 

 the fullgrown spicula (Cf. p. 8). There is 110 striking and constant difference between the 

 sterrospirae of P. carinata and P. meloöesioides. Usually they are more or less kidney- or 

 beanshaped, sometimes almost ellipsoidal ; very frequently restricted in the centre of the 

 "long" axis - - i. e. the axis of the original young stage (spinispire). The size is less variable 

 than in P. melobesioides. 



3. Parenchymal spinispira. (PI. IV, fig. 9 SS — 'CC; fig. io« — 1; fig. II j3 — y; fig. 12 S — y; 

 fig. 13/7 — t). Although by far the great majority of these spicula occur in the parenchyma, a 

 few may be found here and there in the grooves. They can easily be distinguished from the 

 dermal spinispirae (microspirae) by their much larger size and quite different appearance. The 

 spines are very long compared to the axis of the spiculum, and are very variable in shape. They 

 are smooth or spined ; the secondary spines are found either on the whole primary spine, or 

 only at the top. As we observed manifold transitions between the smooth spines and those 

 e. g. in specimen 1500 (PI. IV, fig. 1 2 y) we cannot make specific distinctions on this account. 

 Sometimes we found even in one and the same specimen spinispirae with smooth spines 

 and others with secondary spines, e. g. in specimen 163 e (PI. IV, fig. 9 SS — ££), or spinispirae 

 with secondary spines, placed irregularly, while others possess spines "subdivided near the 

 end into simple or bifid cladi" as Sollas described for P. carinata-, this is e. g. the case 

 in specimen 1458 (PI. IV, fig. 13 vj — 1). 



4. Microspira. (PI. IV, fig. gv — b; fig. 10 B — v; fig. 1 1 d — x ; fig. 12 a). The microspirae are 

 smooth or slightly spined. They are very minute and a very high power is necessary in 

 order to understand their true shape. They vary considerably in size and shape, as can be 

 best demonstrated by the illustrations. Comparing however a large number of them it becomes 

 evident that they are K-spiraxons, which in some cases become very much reduced; though 

 never so far as to form anything like a spherule. Transitions between these spicula, which 

 Sollas called microstrongyles, and those which the same author called spherules we have 

 never met with, although we have paid special attention to them. (Cf. infra p. 15). We agree 

 with Lindgren that Sollas described a part of the young stages of sterrospirae likewise 

 under the name of microstrongyles. The true ones, for which we prefer to take the name 

 microspirae, proposed by one of us (Vosmaer 1902 p. iro), belong to the typical dermalia; 

 they occur as a thin sheet of about two layers at the outer surface of the sponge, distally 

 from the cortical sterrospirae ; this sheet is continuous over the grooves. In the distal portion 

 of the cortex they literary fill up the spaces between the sterrospirae (PI. V, fig. 11). As 

 typical dermalia they line the peripheral canals, forming only a single layer, but a very 

 dense one. In the deeper parts of the sponge they are less abundant in the walls of the 

 canals ; often it seemed to us that the excurrent canals have less microspires in their walls 

 than the incurrent ones. At the sponge-surface as well as in the walls of the canals they 

 are situated with their (longitudinal) axis parallel to the surface or the canal. 



5. Spheraster. (PI. IV, fig. 1 1 a). In our specimens spherasters are rare or absent. It may be 

 that in the latter specimens (1458, 1500) spherasters are still present in other parts of the 



