424 Mr Gardiner , On the Unit of Classification 
in general perceived when a number of specimens be examined, 
and need not be dealt with. The other two classes of characters 
differ in that the one shows discontinuous or specific variability, 
and the other continuous or normal variability 1 . The former 
class serves to define the species, while a scientific unit for the 
second class can only be found in each several specimen. The 
coralla in their characters of the last class intergrade into one 
another in every conceivable way. One or more of these characters 
may become fixed in any set of specimens, and hence of specific 
importance. In the same way characters, specific in most species, 
may become continuously variable in others. Generally, how- 
ever, the interchange between these two groups of characters is 
extremely rare in corals, and affects only a limited number of the 
characters. 
The term variety serves for the different groups in a species, 
which show in their specific characters discontinuous variation. 
The separation of these from species depends on the exami- 
nation of a not inconsiderable number of specimens as well as 
the insight of the individual worker. Their determination requires 
the examination of thousands of specimens rather than tens, 
a task of Herculean proportions in the Madreporaria. If the 
presence in corals of the three classes of characters mentioned 
above be generally recognised, the fault of the systematist can 
only lie in the confusion of species and varieties. Mr Bernard’s 
form , scientifically determined, is almost the same — so far as 
I conceive it from his paper — as a variety. If it be the same, 
it is based on a scientific foundation, but if it is meant to be 
any grade below this, it is surely devoid of such groundwork. 
The only real base of a lower order is the sum of the characters 
of each specimen, the foundation adopted by most systematists. 
It appears to me that the free recognition of the ascertained, 
probable and possible facts of variation would do more to clarify 
systematic biology than the adoption of additional, purely arti- 
ficial standards of other orders. 
I may for a moment return to Mr Bernard’s proposal in its 
aspect of practical utility, leaving for the moment completely out 
of account its scientific bearing. I would ask whether, supposing 
a large collection of corals be not known to the investigator to 
come from the same locality, the forms would be easier to deter- 
mine than in a similar-sized collection from mixed localities. It 
appears to me that the difficulties of ascertaining our forms — 
unless we mean by the term species, varieties or specimens — will 
be even harder than that of determining species , as the latter is 
1 Vide “ Materials for the Study of Variation ” (1894) and “ Heredity, Dif- 
ferentiation, and other Conceptions of Biology,” Proc. Boy. Soc. Vol. lxix. 1901, and 
numerous other papers by W. Bateson, F.R.S. 
