480 ANDREWS: PROTOPLASMIC STREAMING IN MUCOR 
trolled by the dimensions of the glass capillary tube, which is often 
of importance. Small bubbles may in this way be brought to only 
a part of a long filament and the effect studied. A stream of small 
bubbles generally quickly ran together when they touched a 
filament near one another. It was possible by this means to 
expose for a time to dry air, only one side of a filament that had 
previously been completely submerged. The streaming, which 
occurred due to the bubble of dry air forced’ in through the glass 
capillary, was slight if only a small part of the filament was touched 
or covered. If a number of bubbles were let in, so that a larger 
part of the filament’s surface was exposed to dry air, the stream- 
ing was somewhat faster, and it continued to increase according 
to the amount of surface thus suddenly exposed to the dry air. 
In these experiments the streaming was never so fast and did 
not continue so long as in the experiments where the fungal hyphae 
were exposed to moist and then dry air in a gas chamber. The 
streaming was always to the part exposed to the dry air where 
transpiration was taking place. When the filament was resub- 
merged, streaming occurred for a time away from the part that 
had been exposed to dry air, but it was never so rapid as it was 
to the tip when this part wasindry air. These experiments are also 
a confirmation of all those investigations in which it was prov 
_ that in a saturated air in a gas chamber streaming which ha 
begun ceased, and when dry air was let in began again. What 
is practically the same is that streaming may be induced by 
forcing warm dry air by means of a glass capillary tube gently 
among the fungal filaments in a hanging drop, if it be continued 
long enough and if the stream of bubbles be rapid. The streaming 
was in the latter experiment slower than in the preceding exper 
ment. Some of the bubbles were caught in the mesh of fungal 
filaments and held; and in these cases where they touched and 
remained in contact with the filaments streaming was 0U 
to that point, but after a few seconds or minutes it ceased. 
The method of using a capillary glass tube was not at first 
so easily manipulated, and many trials and the loss of a good 
many specimens resulted before it was successful to my entire 
satisfaction. 
; ; iration 
Some other specimens for these experiments on transpira 
