Marsh, and its Relations to Evaporation , &c. 315 
of evaporating more than some 2 c.c. per hour ; but that up to that amount 
it would behave very much like the other two instruments. This evapori- 
meter was therefore kept in the lowest position, C, during the field experi- 
ments. Here the actual evaporation was always so small that the 
instrument was quite capable of meeting the demands made on it. In 
all other cases, whether on the roof, or in the laboratory, the evaporation 
ratios for the different cylinders remained fairly constant. 
It has been shown that, per unit area, the fine-grained porous cylinders 
used in 1908 had an evaporating capacity some 1*28 times as great as the 
somewhat coarser ones of 1907. This is probably due, partly to the finer 
earthenware possessing a larger number of pores per unit area, and partly 
to the greater capillary effect of the finer pores. So long as there is 
equilibrium between the evaporation outside, and the flow of water through 
the capillary pores, so long will the rate of evaporation increase with the 
fineness of the earthenware. But if evaporation should become very exces- 
sive, this state of equilibrium may be destroyed, owing to the counteracting 
of the greater capillary effect by the increased frictional resistance in the 
pores. On the other hand, the viscosity of water diminishes as the tem- 
perature rises. Thus a high temperature would not only increase evapora- 
tion, but would at the same time lessen the frictional resistance due to the 
viscosity of the water, and so decrease the risk of the steady state not being 
reached. In point of fact, it is probable (as seen above) that this state 
of equilibrium between evaporation and the flow of water through the 
pores was reached in all the field experiments recorded in this paper. 
But the case of 1907, II, shows the advisability of testing the individual 
cylinders before using them in exposed positions. 
On the whole, the above considerations, as well as the greater uniformity 
shown by the 1908 cylinders (cf. the calibration data given above), point to 
the distinct superiority of earthenware of a fine, even grain. 
To convert Cubic into Linear Evaporation. 
In this paper I have given the evaporation results in cubic centimetres. 
But as most published observations are recorded in linear measure, it may 
be useful to give multiplication factors whereby my results may be con- 
verted into either centimetres, millimetres, or inches. 
It has been seen that the standard evaporimeter (1908, I) had an 
evaporating surface of 141-2 sq. cm. : and further, that i-o sq. cm. of the 
porous cylinder was equal in evaporating power to 1-43 sq. cm. of open 
water surface. 
Therefore 1-43x141-2 c.c. will evaporate from the whole cylinder 
during the time that it takes i*o c.c. to evaporate from i-o sq. cm. of open 
water surface. 
