Chap. VI. MILLEPORA ALCICORNIS. 293 



and more active {Figs. 4, i k I m n o, and 5). The latter differ chiefly from the 

 larger hydrte, in having tentacles scattered upon the whole length of the stem, like 

 Halocharis ; but instead of being gradually larger from base to summit, the reverse 

 is the case with the tentacles of the small hydrie of Millepora, the lower ones 

 {Fig. 5, g i h f) being the larger, and those near the summit growing gradually 

 smaller and smaller {k I m). The knobs of all these tentacles are chiefly made 

 up of larger lasso-cells {Fig. 5"), the largest of which have a very long thread, 

 remarkable for the enlargement of its spiral band, at a great distance {c) from 

 the bulb {a). 



Whether the diflference in the size of the hydrte is connected with their fertility 

 or not, could not be ascertained, as no medusfe-buds were observed upon either of 

 them. A most interesting point in the history of this genus remains, therefore, 

 still to be traced. It can, however, hardly be doubted that the hydrae must produce 

 medusae of some kind or other, as all the Hydroids do. 



The Corallum. — It is seldom that in dried specimens of the corallum the natural 

 smooth surface can be studied with confidence, on account of the extreme delicacy 

 of the spongiform mass of most recent growth. It is impossible, even with the 

 iitmost care, to handle a specimen without abrading the slender, irregular spiculae, 

 whose points form the horizon over which the soft walls of the animal stretch 

 in a uniform, smooth film. It is, therefore, necessary to study j^erfectly fresh 

 specimens, in order to form a correct idea of the relations of the superficial, 

 spiculate deposits of the animal basis. Under such conditions the whole corallum 

 will appear dotted with round apertures, usually of two sizes (PI. XV. Fig. 8, a b), 

 in numbers and position corresponding to the large and small hydras, which may 

 be found retracted within their cells. Oftentimes the aperture of the cell is com- 

 pletely closed over by the contracted basal walls of the hydra, and in such cases 

 the actual number of cells is disguised. 



The only proper means of removing the fleshy part of the animal, in order 

 to study the corallum, is a potash solution, after which operation it appears, to 

 the naked eye, like a very fine sponge ; but under a considerable magnifying power 

 it presents a very ragged aspect (PI. XV. Fig. 8), esi^ecially about the tij^s of the 

 branches, where the jagged spiculae are less intimately united with each other. 

 From a superficial point of view, the cells of both the large and small .h3'drse 

 appear to have radiate semi-partitions, which, in some instances, are quite regularly 

 disposed, and in the largest cells amount to a dozen in number; but, upon closer 

 inspection, it may be seen that the apparent lamellte are very irregular, laterally 

 flattened projections {Fig. 8, f-m), which have but little depth, and stand at various 

 levels {Fig. 10, c d e f). In some cases a longitudinal section of a cell {Fig. 13) 

 discloses a complete series of false partitions from mouth (a) to bottom (A), whereas, 



