550 
n. RUGGLES GATES AND NESTA THOMAS. 
individiuils will have 15 chromosomes. The origin of such 
mutants is clear. They originate through one of the 
heterotypic chromosomes passing into the same daughter- 
nucleus as its mate, instead of into the opposite nucleus. 
This occurrence was first discovered by one of us (Gates, 
1908a) in the mutant rubrinervis, and has since been 
observed in several other mutants and species. As for- 
Tiierly pointed out, it is made possible by the weak attraction 
between the homologous chromosomes during diakinesis 
and the heterotypic mitosis. We may therefore predict that 
in other genera showing similar cytological peculiarities, 
sporadic mutants of a similar aberrant type will be found 
to occur. 
Whenever this irregular meiotic division occurs in a pollen 
mother-cell, such a cell will, at least in many cases, give rise 
to two lata-producing pollen grains in addition to two 
having only 6 chromosomes. The latter apparently always 
degenerate. Similarly, when such an irregularity occurs in 
the megaspore meiosis, if the 8-chromosome megaspore 
functions it Avill, after fertilisation by a 7-chromosome pollen 
grain, give rise to a lata-like mutant. The frequency of the 
occurrence of this irregular division in 14-chromosome plants 
may determine the frequency with which lata-like mutants 
wdll appear. Moreover, in lata or semi lata when crossed 
with their 14-chromosome parents or when self-pollinated, the 
percentage in which the mutant reappears will depend upon 
the relative number of their 8-chromosome and 7-chromo- 
some germ-cells which function. 
The frequency of the original unequal division in meiosis is 
difficult to determine cytologically, but from the observations 
of Gates it appears to be of the order of 1 per cent. This 
would give about two 8-chromosome pollen grains in 400 or 
0*5 per cent. If the frequency of this irregularity in the 
megaspore mother-cells is the same, about ] per cent, of lata 
mutations should be anticipated (c f . p. 527). The relative 
numbers of the two types of megaspores produced by lata 
may be estimated from the number of lata plants occurring 
