104 G. H. KNIBBS. 
when he affirms that the correction 1:°505 U?/(29) will make ‘ 
Poiseui peri ts consistent. Clearly however the corrected 
pressure for the entrance of the tube gives the value, not for the | 
- section EE’, but ee’, Fig. 1, the distance between these, however 4 
being unknown, and probably somewhat uncertain as to constancy. 
It is however unlikely that it ever reaches two diameters where 
the end is sharply cylindrical or even one where it is conoidal ot | 
rounded off. Couette,! Boussinesq approving his view, with q | 
greater rigour takes into consideration the features of the flow : 
not only at the entrance but also at the point of efflux from the 
tube. He recognises (a) that its length should be reduced for the 1 
distance between the sections EZ’ and ee’; (6) that frictional | 
resistance is suffered by the liquid before it reaches this latter 
section—where parallelism of flow is fully established : (°) that 
the effluent liquid penetrates that in the receiving reservoil, and 
maintains for an appreciable distance its cylindrical form without d 
sensible change: (d) and that the surrounding liquid offers 
frictional resistance to this penetration by the effluent. If lengths 
of tube whose resistances are equal respectively to the resistances : 
(6), and (d) above mentioned, be denoted by /, and /4, and the 
distances He=E'e' and Ff= F'f’ respectively by /1 and J, the q 
corrected length Z, will be ‘ 
hoe be ite tablets c.. ee eae (12) 
Couette argues that the part of the ‘charge’ expended by frictional 7 
resistances in the vicinity of the extremities of a tube ought to be 
at least approximately proportional to the efflux, as is the cai 
with regard to the tube generally, and he therefore expresses 
consequent reduction of the pressure under the form 84 lq/ («h 
so that / will denote the algebraical sum of the terms /, to 12a 
is affirmed by Couette! that this quantity / may be assumed to ys 
always positive : we shall shew however that it appears sometimes 
to be negative and therefore we consider its sign to be undeter 
1 Annal. de Chimie, 6 sér. t. 21, chap. iii, p. 494. . 
2 FF' to ff! Fig. 1. a 
3 Ibid. p. 501. 
4 Ibid. p. 500. : 
