494 Osborn . — Some Observations on the Ttiber of Phylloglossum . 
Experimental work was confined to the laboratory for convenience of 
examination, and because of the uncertainty of the season. The 1917 
growing season was a wet one, 1 especially during the critical months 
September-October, when most active growth in tuber-production occurs. 
In 1916 also there had been more than average rainfall, but there was no 
reason to expect a continuance of the conditions in 1918, as indeed proved 
to be the case. In the different experiments the aim was usually not so 
much to reproduce the conditions of humidity that obtain in the field, as to 
maintain a constant moisture. It is not suggested that an atmosphere so 
near saturation would be maintained in the field for long together, though 
in some years, e.g. 1916, a high rain- 
fall occurred during the last three 
months of the year, and conditions 
might remain favourable to tuber pro- 
duction over considerable periods. 
In each experiment freshly de- 
tached leaves were laid on soil from 
the locality, kept moist and covered 
by glass. In the first experiment the 
soil was lightly tamped down in a 
Petri dish, but in the subsequent series 
a sod several centimetres square in 
which Phylloglossu m plants were grow- 
ing was used. This method was more 
successful, as will be seen from Series B 
and C, in which 8 and 10 leaves gave 
respectively 7 and 8 positive results, 
but the first method gave a valuable 
series of abnormalities, discussed later. 
In addition to the five series of 
experiments, odd leaves were injured 
but not detached from plants kept in 
the laboratory, the object being to 
repeat the conditions under which such plants as those shown in Text-figs. 
13 and 14 grew. Text-fig. 15 shows the result of such an experiment. One 
of the three leaves of a plant growing undisturbed in a sod of soil was 
severely injured but not detached. The plant was dug up and sketched 
after nearly four months’ growth. It had formed a new tuber in the normal 
way, and the injured leaf, too, had produced a tuber sunken on a short 
dropper below the surface of the soil. Text-fig. 15 should be compared 
with Text- fig. 14, which shows a plant collected in the field. The results 
are closely similar, except that in the experiment the leaf-borne tuber had 
1 See p. 508 below. 
Fig. 15. Three-leaved plant grown in 
laboratory. One leaf, injured July 17, 1918, 
has developed an adventitious tuber. * Marks 
the wounded surfaces which are contiguous in 
specimen (Nov. 10, 1918). x 3J. 
