The British Association at Birmingham. 355 
solution. Pure ammonium humate was obtained by abstracting 
bacteria-treated peat with distilled water, precipitating humic 
acid from the extract by hydrochloric acid, and after repeated 
washing of the humic acid adding just sufficient ammonium 
hydroxide to dissolve it. This was then diluted with distilled water 
in the proportion of 1 humate to 200 water. 
Mr. W. Neilson Jones gave an account of some investigations 
on anthocyan formation, with special reference to the two following 
points. (1) When coloured petals of Matthiola, etc., are soaked in 
95 per cent, alcohol they become colourless, when transferred to 
water they regain their original colour. It is believed that a 
pigment-producing mechanism and a reducing body are present in 
the petals. The amount of water in the cells determines which 
way the reaction shall go: in strong alcohol reduction occurs and 
the petals become colourless; in weak alcohol or water oxidation 
takes place, resulting in the production of pigment. It appears 
further that considerable quantities of reserve “ raw material ” 
occur in some coloured flowers, e.g., Matthiola , from which pigment 
can he produced. The darkening of flowers on fading is explicable 
on this hypothesis as being due to this raw material coming into 
action. (2) Benzidine gives a blue colour with a solution of oxidase. 
Methyl quinol gives no reaction with this oxidase solution alone, 
but on the addition of a trace of benzidine becomes oxidised to a 
red colour. Methyl quinol thus behaves towards benzidine as an 
epistatic member of a colour series (e.g., in Primula sinensis) to a 
hypostatic member—that is, it is able to manifest itself only if the 
hypostatic member is present. In this reaction the benzidine 
probably plays the part of a carrier of oxygen to the methyl quinol. 
Although it is dangerous to argue from analogy, in default of other 
evidence it would seem probable that the colours of an epistatic 
series are due to different substances behaving like the methyl 
quinol, rather than to the action of specific oxidases. 
Dr. R. Ruggles Gates discussed certain cytological evidence 
to show that mutation and Mendelian splitting are different processes. 
Definite evidence has been obtained showing that some at least of 
the mutations of Oenothera are not due to recombinations of 
Mendelian characters, as various writers have assumed, but to 
irregularities in meiosis which lead to changes in nuclear structure. 
The only cases cited in the paper were those of O. lata and 
O. semilata, which offer a means of differentiating between 
characters inherited from the parents and those resulting from 
unequal or irregular chromosome distribution during meiosis. 
Miss Lutz and the author showed independently that the mutant 
lata has 15 chromosomes instead of 14, and along with 
Miss N. Thomas the author has recently found that the same is 
true of semilata, also that all individuals possessing the foliage and 
habit of lata or semilata contain 15 chromosomes, even when these 
characters are associated with others derived by inheritance from 
their parental forms. Thus in a mutating race of 0. biennis 
from Madrid, one mutant was O. lata biennis, having lata foliage 
and biennis flowers; this plant has 15 chromosomes, and must have 
originated through a meiotic irregularity. Again, in F 2 of 0. rubri - 
