380 Knight . — Further Observations on the Transpiration , 
discussed at length elsewhere by various writers and need not be further 
considered here, but it is necessary to refer to the work of Livingston and 
Brown, which has a bearing on the phenomena affpresent under considera- 
tion. In their carefully controlled experiments, the difference between the 
maximum and the minimum leaf water-content was never more than 8 per 
cent., and in one case, Physalis, only 1 per cent., with an average for all re- 
corded cases of 5 per cent. This figure, obtained under extreme atmospheric 
conditions, is widely different from that obtained by Clark, but is probably a 
much more accurate estimate, as the experiments recorded below tend to show. 
In view of the very small decrease in leaf water-content which will produce 
• flaccidity, it is evident that if during the normal daily cycle the leaves of 
the plants concerned exhibit fluctuations of water-content such as those 
recorded for the desert habitat, the plants must daily suffer a considerable 
degree of wilting. This is not the case ; on the contrary, the plants under 
consideration do not wilt to an extent appreciable by observation except in 
a high wind or as the result of deficient soil moisture. Consequently it was 
thought desirable to apply the method of Livingston and Brown to the 
species used in the present work in order to determine the extent of the 
diurnal changes of leaf water-content. The plants were living in a cool 
greenhouse or under temperate outdoor conditions, not approximating to 
those of the desert habitat of the plants of Livingston and Brown. Experi- 
ments were carried out with ash, apple, and quince out of doors and with 
Eupatorium, Peristrophe , and Cycas in a greenhouse. In the case of apple 
and quince large numbers of leaves were taken and the probable errors were 
calculated. In dealing with the other species companion leaves or leaflets 
were compared in the manner indicated in experiments recorded above. 
The period of maximum water-content of the leaves used by Livingston 
and Brown occurred during the morning hours, whilst the minimum was 
reached during the afternoon. With this result in view determinations were 
made in the present experiments at two different times of the day ; for 
example, at 8 a.m. when the atmospheric evaporating power was low and 
the loss by transpiration small, and again at 2 p.m. after a period of rapid 
transpiration. Table VI gives the result of an experiment using pairs of 
leaves of Eupatorium adenophorum. 
It has previously been emphasized in this paper that in experiments on 
the water-content of pairs of leaves it must be recognized that the accuracy 
is limited by the extent of the initial difference between -the water-contents 
of the two companion leaves. The last two columns of the table show that 
the evaporating power of the air, as determined by a Livingston porous cup 
atmometer, was at its highest during the morning, between the 8.30 a.m., 
and the 1 p.m. water-content determinations, whilst during the period 
between the 1.1 5 p.m. and the 5*15 p.m. determinations it had fallen to a 
quarter of its morning value. 
