April 27, 191 1] 



NATURE 



279 



end of the lower half of the lake. The water, however, 

 \yas shallow in this part of the lake, and the records are 

 rather disturbed by local effects due to wind. The periods 

 of the seiches detected here were 204 minutes, 11-9 minutes, 

 and 33 minutes. The 20-4-minute and ii-g-minute 

 periods are due to the uninodal and binodal seiches of the 

 lower half of the lake. The 3-3-minute period is probably 



Fig. I. — Recording Limnograph. 



due to a transverse seiche. The maximum range recorded 

 was about an inch. 



After this, the recording apparatus was set up at the 

 head of the lake. The upper half of the lake being much 

 deeper than the lower half, better records were obtained. 

 The maximum range recorded was about \\ inches. The 

 periods here were well marked, and had the following 

 times : — 697 minutes, due to the uninodal seiche of the 

 whole lake; 14-1 minutes and 6-6 minutes, due to the 



and 3 are typical traces obtained at the head of the lake. 

 Nos. 4 and 5 are obtained from No. 3 by Prof. Chrystal's 

 method of residuation. The 3-4-minute period is seen in 

 the original trace. No. 4 shows the i4-i-minute and 

 69-7-rninute periods, and is obtained by residuating out the 

 3-4-minute and 6-6-minute periods. No. 5 shows the 

 Ob-minute and the bg-y-minute periods, after the 3-4-minute 

 and the i4-i-minute periods have been residuated out. In 

 Nos. 1 and 2 the rate of movement of the paper was one 

 inch in i8-8 minutes, and in No. 3 one inch in 232 minutes. 



During the later experiments a form of apparatus was 

 used which proved quite satisfactory, and, being simple to 

 construct, may be worth briefly describing. A strip of 

 paper, from a continuous roll fixed on the base of the 

 instrument, passes up and over a horizontal wooden 

 cylinder, 3 inches in diameter, and driven by clockwork. 

 After passing half-way round the cylinder, the paper passes 

 under a small roller carried on springs. This roller presses 

 the paper against the wooden cylinder, and, since the paper 

 passes half-way round the cylinder before passing under 

 the roller, there is no possibility of it slipping. 



A horizontal lever is pivoted to the base of the instrument, 

 one end of which projects outwards, and is connected to the 

 lloat by a string, while the other end carries a weight. 

 The pen and holder are carried by a horizontal rod, which 

 is supported by two upright arms, being fixed to them at 

 each end by pivots. One of these arms is fixed to the 

 lever, at the place where it is pivoted to the base of the 

 instrument, while the other arm is pivoted direct to the 

 base. As the float moves up and down, this horizontal rod 

 moves backwards and forwards, parallel to the axis of the 

 wooden cylinder. On the horizontal rod are bearings, 

 which carry the light frame holding the pen, which rests 

 on the top of the wooden cylinder. When the lever is 

 half-way up or down, the bearings of the pen are about the 

 same height as the top of the wooden cylinder, then, as the 

 float moves the lever up and down, the pen moves in an 

 almost straight line across the paper on the top of the 

 cylinder. 



This apparatus is simple to construct, and, since the only 

 friction is in the pen and the four pivots, the whole system 

 moves very freely, and a float 5 inches in diameter will 

 work it easily, while Chrystal's " waggon " recorder 

 requires a lo-inch float. Gordon Dobson. 



Caius College, Cambridge, April 19. 



I ^Ayvv-vwAvn^rv^'^^^ 



A^^'^-^y^^/^^-^'^Ayv'v''"^'VVvyv^'^^^ 



Fig. 1. — Traces obtained at Head of Windermere. 



uninodal and binodal seiches, respectively, of the upper half 

 of the lake ; also a well-marked, short period of 34 minutes. 

 This last period is probably due to the trinodal seiche, and 

 also to a transverse seiche, which has nearly the same 

 period. In some cases, the movement due to this oscilla- 

 tion alone was nearly an inch. 

 The figure shows some of the traces obtained. Nos. 1,2, 

 NO. 2165, VOL. 86] 



The Flight of Exocoetus. 



Practical difficulties will prevent the settle- 

 ment of the question as to whether or no a flying 

 fish supports itself by movement of its fins by 

 the method suggested by a correspondent in 

 Nature of February 9, viz. kinematograph photo- 

 graphy. 



Anatomy and phylogeny converge to the sup- 

 port of those observers who declare that the 

 " wings " are motionless during " flight." 



(i) Any resemblance to the huge musculature 

 of birds is out of the question, but if the wings 

 vibrate to any purpose, something resembling in 

 scale the muscular and nervous specialisation 

 found in insects should obtain here. Has any- 

 thing of the sort been found? On the contrary, 

 the muscular development of Exocoetus is, like 

 that of other fish, directed to propulsion by the 

 tail. 



(2) The structure and habits of the lower 



members of the family, Hemiramphus and Belone, 



indicate stages in the evolution of Exocoetus. 



The former is able to make great leaps nearly 



parallel to the surface, of such force, indeed, that 



the natives here tell me of men who have been 



_— — -— pierced by the elongated lower jaw two inches 



deep in the flesh of the leg when wading among 



them. " When it is out of the water it is quite 



mad and strikes whatever is in the way, whether 



a man or a boat, and so kills itself," to quote their 



description. 



Belone can almost fly, its effort having the appearance 

 of running on lh«> surface on the tip of its tail, suggesting 

 some use of this mombor, but not of the normal-sized fins, 

 in extending tho range of " flight." These two stops in the 

 evolution of the habit of Exocwtus distinctly lead to the 



