1898.] SPECIES OF iriLLEPOiiA. 247 



increases it becomes more evident that in others uo satisfactory 

 classification can be framed until we have a thorough knowledge 

 of the anatomy of the polyps which construct these skeletons and of 

 the canal-systems which bind them together into colonies. 



In some genera of Madreporaria, for example, of which the 

 skeletal chfiracters only are known, a long series of intermediate 

 stages can be found between the type specimens of the different 

 species, and every new collection of specimens that is examined 

 increases the difficulty of deciding whether a particular inter- 

 mediate form belongs properly to one species or another. Moreover, 

 in this same group the outlying species of one genus resemble the 

 outljdng species of another so closely that it is often a matter of 

 great difficulty to determine, on our present system, to what 

 genus a particular specimen belongs. 



Nearly every important systematic work on these Coelenterates 

 contains some remarks about the difficulty of determining species, 

 and examples are quoted of series of intermediate forms con- 

 necting closely allied species. If it were possible to frame some 

 general rule for the correct definition of a species, which Avould 

 be agreed to by all systematic zoologists, our task might be less 

 difficult than it is ; but, as matters stand, the conception of what 

 is a species of one worker is so different from that of another 

 that there is constantly going on a see-saw of construction and 

 destruction of new species in our systematic literature. 



I do not propose to attempt to define the conception "species" 

 in Coelenterates, but I think that all zoologists would agree that, if 

 a form which is known as species A were proved to give rise to an 

 embryo which grew into a form which had hitherto been known 

 as species B, the two forms would have to be merged into one 

 species with one specific name. Similarly, I imagine that all 

 zoologists would agree that if a coral known as species X changed 

 in the course of its life-history into a form known as species T, 

 then the forms X and T should be regarded as one species and 

 retain only one name. In the absence of any experimental proof 

 that the embryo of one so-called species of coral gives rise, imder 

 any circumstances, to another so-called species, or that one 

 so-called species changes in the course of its life-history into 

 another, it is necessary to examine with very great care the 

 anatomy of the soft parts as well as the skeletal structures, in 

 order to determine whether it is possible or even probable that 

 such changes actually occur in nature. If we find, then, that the 

 polyps or reproductive organs of a coral with one form of growth 

 are essentially different from those of another form, we may 

 consider there is good reason for believing that such changes do 

 not occur and the species founded on the skeletons are good ; but 

 if, on the other hand, the polyps, reproductive organs, and other 

 characters of the two forms are essentially the same, then there is 

 reason for believing that the species founded on skeletal characters 

 may not be good. 



Before proceeding farther with this discussion of the characters 



