140 KEPORT— 1898. 



difficulty in Jetecting minute variations, and no hesitation or doubt as to 

 when accurate agreement is obtained. 



Both water and alcohol have been tried in a similar test case, although 

 the conditions are much more difficult to comply with in these cases, thin 

 sheets, such as ^-f,\p^,th of an inch, which have to be used, making the 

 experiments much more difficult. 



The results of stream- line flow by this method may therefore be 

 received with confidence, and a number of cases of stream lines have been 

 obtained by using glycerine. These experiments were made in the first 

 case with the view of studying discontinuity under conditions which 

 involve a severe test of the stability of the thin film, which was through- 

 out of a thickness of viViT^h of an inch, and in all cases forty-one colour 

 bands were used within a width of about 3 inches. Without the figures 

 themselves, which, as already mentioned, are published elsewhere, it is 

 impossible to do more than describe the general results ; but it may be 

 said that the difference between using square corners and rounded edges is 

 very marked ; indeed, it was a matter of surprise to find the flow main- 

 tained so well over the sharp edge at all. It is evident that the liquid 

 which actually adheres to the edge of the obstacle enables a definite 

 though very minute rounding to take place at the corners, as is visible by 

 a close examination of the photographs. This was especially marked in 

 one case on one side of an experiment, whereas on the other side, when 

 the entering colour band actually touches the edge, the sharp corner takes 

 effect upon it, and completely breaks it up, destroying the continuity of 

 flow. 



It may be said that in these examples the narrow portion of the orifice 

 is so small that it is impossible to detect the separate bands which pass 

 through that portion ; these colour bands, nevertlieless, emerge quite 

 distinct from each other, and finally assume their original position in the 

 wider portion of the channel. One example may specially be mentioned, 

 viz., that of a flat plate inclined at 4-5° to the stream. The agreement of 

 this case with the theoretical result of Professor Lamb the author has 

 previously been able to verify in the case of water, but even with the 

 greatest care it was always possible to tell in which direction the stream 

 was flowing. With glycerine, however, the colour bands are practically 

 identical before and behind the plate ; and if it were not for the point 

 being clearly evident at which the central divided band reunites, it would 

 l:ie impossible to tell which way the stream is flowing . This point of 

 union of the two portions of the central band is extremely interesting, as 

 a careful measurement of it verifies the exact position at which the central 

 stream line meets the plate, and is found to agree precisely with the 

 mathematical solution of the problem. 



(6.) Method of Inrestiyating Effects of Variable Resistance. 



In order not to prolong the present paper beyond reasonable limits, 

 the author will only briefly mention a method by which variable resistance 

 can be dealt with. If within the thin sheet of flowing liquid an obstacle 

 of some transparent material less in thickness than the sheet itself be 

 placed, the flow will take place partly over the obstacle and partly round 

 it. This effect will correspond to that of an obstacle through which fluid 

 can flow, but which opposes resistance greater than the remaining portion 

 of the thin sheet. 



