138 REPORT— 1898. 



foil, or other materia] a border, together with the required obstacle or 

 channel, and clamping the whole together, an effective and simple lantern- 

 slide is obtained. 



(3.) Result of using Liquids other than Water, 



The four liquids other than water which were experimented with were 

 castor oil, glycerine, alcohol, and mercury, of which glycerine is so entirely 

 and surprisingly satisfactory in every respect as to make it undoubtedly 

 the best material which the author had hitherto worked with for obtain- 

 ing stream-line figures, and the whole results shown at the reading of 

 the present paper were obtained by using glycerine. 



Of course after the glycerine has once passed through the lantern 

 slide the coloured portion has mixed up with the clear, and it can only be 

 employed again for colour bands, but at the same time the thickness of 

 the sheet of flow being small, while the velocity with which perfect results 

 can be obtained is low, there is no reason why this material should not 

 be always employed. 



Alcohol has such a low viscosity that it can be employed in sheets of 

 remarkable tenuity ; but these sheets have naturally the disadvantage of 

 giving colour bands so thin as to be scarcely capable of photographic 

 reproduction, while the volatile nature of this substance makes it not 

 altogether desirable in close proximity to an arc lamp. 



Mercury is of course opaque, but its great density compared with its 

 viscosity makes it most valuable in connection with some experiments, 

 and its lines of flow can be traced if it is not quite pure by marks it leaves 

 on the glass. It is, however, troublesome, since it cannot be used in con- 

 nection with brass taps or with the brass lantern apparatus. 



Castor oil also proved troublesome to work with. 



(4.) Measurements made to co}npare the Floiv with Water, Glycerine, and 

 Alcohol in Test Cases, and Explanation of these Hesults. 



In the test cases above referred to Professor Lamb kindly sought for 

 and obtained an equation for the stream lines round a cylinder in a 

 parallel channel, and the results of the measurements, although warranting 

 the use of water under these conditions, showed that the flow was not in 

 absolute agreement with the lines plotted from the formula. The author 

 then remarked: — -'Although the differences are appreciable, they are to some 

 extent of a nature which must be attributed to the great difficulty, in the 

 first place, of making sufficiently accurate mechanical arrangements, and 

 also from the fact that it takes some little time to plot down the results ; and 

 that, during this time, it is extremely difficult with the present appliances 

 to keep a ferfectly steady and uniform pressure both of the colouring 

 bands and the main body of the water, when each comes from a separate 

 source. Beyond this, however, there is no doubt that the stream lines are 

 slightly pushed away from the obstacle at the point of greatest velocity, 

 i.e., abreast the mid-section. This may be due to the slight effect of 

 viscous resistance parallel to two containing glass boundai-ies.' 



In view of the importance of the matter it seemed worth while 

 to attempt a new and more accurate comparison of the experimental 

 results with the flow for a perfect fluid. In the previous case the 



