(ENOTHERA ML’T. LATA AND CE. MUT. SEMILATA. 501 
Tliere are two inetliods of studying tlie nature of any par- 
ticular mutational change. One is the comparison, cytologi- 
cally as well as morphologically and chemically, of the mutant 
with its parent” the other is by studying the inheritance of 
the mutation in its offspring and in crosses with its parent 
race. When both these methods are used together they sup- 
plement each other, and the cytological investigation of 
mutants in particular has served as a useful check on 
Mendelian speculation which would explain all mutants in 
.terms of a single idea. 
In the following table I have attempted to classify the 
various types of mutational change. Though the classifica- 
tion is necessarily incomplete, it serves to indicate the varieW 
of types of mutation which are now known. Examples of 
each type are given from the many instances available in the 
recent literature. The cases have been selected both from 
plants and animals, and the name of the organism with the 
name of the discoverei- or investigator of the mutation is given 
in each case. 
Near the borderland between mutations and fluctuations, 
and partaking somewhat of the nature of both, are such 
variations as variegation of foliage (e. g. A.cer striatum 
variegatum Godron), and striped flowers which, according 
to Yilmorin, originate through partial reversion from white 
varieties. The ever-sporting varieties described by 
de Yries are also perhaps to be placed here. 
Dobell (1913), in his recent useful review of the work on 
mutation in Trypanosomes and Bacteria, uses the term 
mutation to denote those heritable modifications which have 
been induced in various ways in various micro-organisms.’^ 
This appears to be essentially the same as my conception of a 
germinal change.” Again, by mutation in Bacteria he 
understands that in a given race individuals may occur 
which differ from their fellows in their genetic constitution.” 
In certain cases these differences seem to be accpiired not 
suddenly, but by several small successive steps. 
It will thus be seen that a host of phenomena of widely 
