1912] on the Total Eclipse of the Sun, April 1911. 



357 



throw an image of the sun on a previously-marked disk (see Fig. 3, 

 right-hand side). The face of this disk was so graduated as to enable 

 the observer to estimate the angle subtended at the centre of the 

 dark moon by the remaining bright crescent of the uneclipsed por- 

 tion of the sun ; previous calculations had shown that when the 

 crescent subtended angles of 90°, 45° and 30°, there remained 42, 9 

 and 4 seconds respectively before totality began. This method only 

 holds good when clouds do not obstruct the view of the sun. 



In order to allow for the contingency of second contact not being 

 seen on account of clouds, the observer was furnished with a deck- 

 watch to give all the necessary signals at their computed times. For 

 the actual calling out of the time during tlie eclipse, a special " eclipse 

 clock " was taken out. This clock has a seconds pendulum, and a 

 special dial is made for each eclipse. At the commencement of 

 totality the clock is set going by releasing the pendulum held by a 

 thread, and the hand (there is only one) makes one revolution a 

 minute. The dial is marked in the form of a spiral, and each second 

 of the duration of totality is marked on it. The time called out by 

 the timekeepers is always the number of seconds remainng before 

 third contact, or the end of totality. On the present occasion totality 

 was computed to last for 217 seconds, so that the timekeepers called 

 out "217," "215," "210," "205," and so on until 10 was reached, 

 when they called out each second. The duty of the observer at the 

 cusp telescope was to be responsible for giving all the time signals 

 up to the beginning of totality, and at his signal " go," the time 

 callers took up the duty of calling out the time. Two men were 

 employed at the clock, one to call the first half and the other to call 

 the second half, so that each man might have an opportunity of 

 seeing the eclipse. The actual code of signals was as follows : — 



10 minutes before totality (wind clocks, 

 caps oS, lamps lit) 



5 minutes before totality 



42 seconds ,, ,, 



9 „ „ „ 



4 „ „ „ 



Totality begins, " 217 " on eclipse clock 

 Totality ends, " " on eclipse clock 



EcHpse day, April 28th, or, as we had not altered our date, since 

 we crossed the "date Hne," April 29th dawned. It was a cloudy 

 morning, there was a considerable amount of roll cirrus and cumulo- 

 nimbi were not absent. In fact it was much too cloudy to make one 

 believe that it would clear up for the eclipse. 



The order of the day was for all parties to be at their stations 



2 B 2 



