176 DISTRIBUTION. 



the JEginidce he appears to have good grotinds. Its enonrious size, however, when compared 

 with living representatives of this family, constitutes a striking feature in its physiognomy. 



Several other fossil impressions of medusae have been described by Ilaeckel, though none of 

 these can with anything like certainty be referred to the Hydroida. Some of them are pre- 

 served with great distinctness, and are undoubtedly referable to the Ducophora or steganoph- 

 thalmic medusas, while some others are so imperfectly preserved as to I'ender it impossible to 

 determine their systematic position, though in some, at least, the evidence is in favour of their 

 hydroid rather than their discophorous affinities. 



It wiU be seen that the only two fossil hydroid medusae which have been satisfactorily deter- 

 mined appear to belong, one to the Truchjnemidce, and the other to the ^E(jinid(B, families charac- 

 terised by certain anomalous features which distinguish them from the ordinary hydroid medusae. 



GrapioVdes. — Among the extinct forms of life few possess more interest than these remark- 

 able fossils, absolutely confined, as they are, to one great section of the palaeozoic rocks, where 

 their vast abundance, wide geographical distribution, and easy recognition, render them of special 

 value to the practical geologist. 



The graptolites are now by most palaeontologists refen-ed to the Hydroida, and their 

 living representatives are sought for among the calyptoblastic genera of this order. While, 

 however, I am unable to recognise their hydroid relations from the point of view from which 

 palaeontologists have generally agreed to regard them, I believe that their affinities with the 

 Hydroida are too decided to justify their omission from any complete exposition of the palaeonto- 

 logical history of this group of the animal kingdom. 



The typical form of a graptolite is that of a narrow tube, straight or more or less curved, 

 emitting from one side a series of hollow denticles, through which the cavity of the tube opens 

 externally, and having a solid slender rod imbedded in the walls of the opposite side. This 

 type form (" monoprionidian ") is represented by the genus Graptolites proper, where the 

 denticles or tubular offsets from the common canal are in contact with one another at theii' bases 

 and usually for a greater or less extent of their length, and by the genus JRasfrites, where 

 they are separated from one another by considerable intervals. 



But we may conceive of two such graptolites being united back to back, and the resulting 

 form will then present two series of tubular offsets, one on one side of the main tube and the 

 other on the side diametrically opposite, while the solid rod will now occupy the axis, holding 

 just such a position as it would do if it had been formed by the union of the two rods of the 

 component halves. 



This form (" diprionidian") is represented by such genera as Dijjiofjraptiis, where the tubular 

 offsets stand out more or less free from the sides of the main tube, and by CUmacograptus, where 

 they are adnate to one another, so as to appear entirely immersed in its walls. 



Some other forms also exist, such as Dicrano(jraptm, in which the graptolite, with a double 

 row of denticles, after continuing its course for a time, divides into its component halves, which 

 then diverge from the basal portion as two branches, constructed each on the single-rowed type. 

 Branched single-rowed forms {CIado(/raptus, Dichoi/rapfu-s) also occur. In Bkhoyraptus priuiary 

 branches radiate from a common point at the proximal end, where they are connected by a web- 

 like disc, apparently com[)osed of a double membrane of the same nature as that which forms 

 the walls of the branches.' 



' See Hull, 'Graptolites of tlie Quebec Group.' 



