HirncENDoRrF.— Natural Self-fertilization of Wheat on a Large Scale. 575 
of the hypothesis. But the seed from this crop should show extensive 
variation. If any particular head were an impurity all its seeds should 
produce like plants; if it were a cross-bred its seeds might be expected to 
produce plants varying inter se.” This was tested by taking twenty-five 
heads of the most varying appearances from each of the three strains and 
owing them in head-to-row plots. At the time of writing,* these plots 
are just ripening, and it is apparent that the variants selected last year 
were of three different classes :— 
(a.) Accidental impurities—in which all the plants of the row are alike 
but are not Tuscans of the type of the plot from which they were 
selected. 
(b. Variants due to differences in soil or situation—in which all the 
plants of a row are alike and belong to the strain from which they 
selected. 
(c.) Natural erosses—in which the plants in a row vary infer se. In 
most of the cases the crosses are with other types of Tuscan, 
but in three rows the cross is with another variety that was 
grown in the same field. 
The following table shows the number of each class of variants selected 
from each of the three strains :— 
E 
Strain. Impurities, [Non-congenital) Crosses | Totals. 
G2 es iu 3 2 20 25 
F8 22 HE 3 13 9 25 
F10 = ER c2 10 11 25 
heterogeneous, and this is reflected in the large number of proved crosses. 
train F8 a small number of crosses appears, and large a number of the 
variants selected were due merely to external conditions. This is re ected 
in the rather remarkable fact that, of the twenty-five rows of this strain, 
the heads selected were of one strain—#.e., the differences were accidental 
— the conclusion independently reached by an examination of the rows. 
) The third method of testing the hypothesis was to attempt to repeat 
the conditions and see if the supposed results were again obtained. The 
minimum temperature on the grass at Lincoln on the night of the frost 
* The completion of the paper was delayed for some weeks to allow almost complete 
ripening of the wheats under observation. 
