TEMPERATURE OBSERVATIONS IN LOCH EARN. 



761 



TABLE V. 



Row 1 



168 



166 



217 



423 



588 



552 



535 



475 



207 



137 



2 



88 



108 



246 



243 



344 



474 



581 



673 



729 



750 



3 



791 



742 



701 



578 



550 



557 



569 



534 



582 



534 



4 



448 



400 



346 



167 



86 



70 



96 



173 



330 



431 



5 



609 



834 



866 



894 



962 



941 



922 



871 



851 



842 



6 



791 



748 



715 



667 



607 



542 



479 



409 



356 



296 



7 



227 



204 



290 



502 



558 



687 



808 



727 



689 



679 



8 



619 



611 



670 



612 



635 



661 



641 



595 



585 



578 



9 



483 



447 



432 



411 



392 



379 



398 



408 



545 



606 



10 



600 



599 



550 



430 



453 



454 



440 



472 



500 



494 



11 



512 



478 



403 



389 



359 



319 



349 



353 



349 



359 



12 



372 



395 



449 



493 



522 



469 



431 



379 



381 



336 



13 



330 



356 



364 



355 



377 



363 



335 



333 



299 



318 



14 



359 



382 



466 



554 



572 



624 



610 



586 



585 



620 



6397 



6560 



6715 



6718 



7005 



7092 



7194 



6988 



6988 



6980 



The largest divergence is from 6397 to 7194, which gives an amplitude of 797. 

 This, divided by the number of "laps" used — i.e. the number of rows — gives as 



"reduced amplitude" = 57. 



14 



TABLE VI. 



Rowl 



168 



166 



217 



423 



588 



552 



535 



475 



207 



137 



88 



108 



246 



243 



344 



474 



581 



673 



729 



750 



2 



791 



742 



701 



578 



550 



557 



569 



534 



582 



534 



448 



400 



346 



167 



86 



70 



96 



173 



330 



431 



3 



609 



834 



866 



894 



962 



941 



922 



871 



851 



842 



791 



748 



715 



667 



607 



542 



479 



409 



356 



296 



4 



227 



204 



290 



502 



558 



687 



808 



727 



689 



679 



619 



611 



670 



612 



635 



661 



641 



595 



585 



578 



5 



483 



447 



432 



411 



392 



379 



398 



408 



545 



606 



600 599 



550 



430 



453 



454 



440 



472 



500 



494 



6 



512 



478 



403 



389 



359 



319 



349 



353 



349 



359 



372 1 395 



449 



493 



522 



469 



431 



379 



381 



336 



7 



330 



356 



364 



355 



377 



363 



335 



333 



299 



318 



359 382 



466 



554 



572 



624 



610 



586 



585 



620 





3120 



3227 



3273 



3552 



3786 



3797 



3916 



3701 



3522 



3475 



3277 3543 



3442 



3166 



3219 



3294 



3278 



3287 



3466 



3505 



The largest divergence is from 3120 to 3916, which gives an amplitude of 796. 



This, divided by the number of laps used, is 



796 



= 114. 



When the reduced amplitudes are plotted against the corresponding trial periods 

 the result is as shown in fig. 14. This shows two distinct maxima at about 10 hours 

 and at about 19^- hours — a good agreement with the period as calculated. 



It will be noted that the maximum showing the binodal period is much sharper 

 than the uninodal maximum. This is easily explained when it is remembered that 

 the binodal period is defined by about 12 periods of observation, while the uninodal 

 period is defined by only 5. To use the well-known analogy of the diffraction 

 grating, it is as if there had been used for the determination of a wave-length 

 a grating of double the number of lines. That this is the explanation is seen from 

 the small dotted curve on fig. 14 at period 10. This was derived from the first 



