182 Proceedings of Royal Society of Edinhuo^gli. [march 5 , 
differently. It turned at once to the right and the rotation was 
thus left-handed (diag. 3 &'), or opposite to its previous direction. 
Twelve rotations were recorded, and indeed a thirteenth, all 
keeping steadily in this direction. It was observed twice after- 
wards like the left, and there was no change of the left-handed 
direction, but instead of moving quicker than at first, it moved 
exceedingly slow. 
Owing to these discrepancies, it became necessary to find out, if 
possible, which was the primary and normal direction for each 
palp reversed. Accordingly several were tried, and their first 
direction noted, with the following result : — Left inner reversed, 
right-handed (on base) ; and right inner reversed, left-handed. 
Rate of Movement. — Numerous continuous observations were made 
over extended periods of time. It generally happened that the rate 
was slow at first, then gradually quickened, attained its maximum 
speed, and finally declined. The greatest speed attained was found 
to be a complete revolution in If minutes. It is hardly possible to 
take the maximum and minimum speed, and determine the mean, for 
now and again the palp will stop, and after a short interval resume, 
so that there is not always continuous movement throughout. But 
at that stage, when there is a regular constant rotatory movement, 
without the complication of to-and-fro movements, a fair average 
rate may be struck for that period, to be called the 'partial average. 
Two averages will thus sometimes be given — a general average, in- 
cluding the rotations from the very commencement; and a partial 
average, only extending over a limited and selected number of 
rotations. 
For determining rate of movement, a right and left inner palp 
were selected from the same mussel, and placed together in the 
same plate, under the self-same conditions. Although placed under 
similar conditions, the two did not behave alike, as the results 
will show. 
Left. — For fifteen recorded rotations, the slowest was 17 minutes, 
the quickest minutes, and the average 6 minutes. The first 
revolution took 11 minutes, and the last (recorded) 17 minutes. A 
partial average, including from the 4th to the 12th round, when 
the rate was comparatively regular, gave 3 minutes per round. 
After the 15th round, the movements became very irregular. 
