DETERMINATION OF DEFLECTION OF LIGHT BY THE SUN'S GRAVITATIONAL FIELD. 313 



climate is very moist, but not unhealthy. The vegetation is luxuriant, and the scenery 

 is extremely beautiful. We arrived near the end of the rainy season, but the gravana, 

 a dry wind, set in about May 10, and from then onwards no rain fell except on the 

 morning of the eclipse. 



We were advised that the prospects of clear sky at the end of May were not very 

 good, but that the best chance was on the north and west of the island. After in- 

 specting two other sites on the property of the Sociedade d'Agricultura Colonial, we 

 fixed on Roa Sundy, the headquarters of Sr. CARNEIRO'S chief plantation. We were 

 Sr. CARNEIRO'S guests during our whole visit, and used freely his ample resources of 

 labour and material at Sundy. We learnt later that he had postponed a visit to Europe 

 in order to entertain us. We were also greatly indebted to his manager at Sundy, 

 Sr. ATALAYA, with whom we lived for five weeks ; his help and attention were 

 invaluable. Mr. WRIGHT and Mr. LEWIS of the Cable Station kindly assisted us as 

 interpreters when necessary. 



Sundy is situated in the north-west of the island overlooking the sea at a height of 

 500 feet, and as far as possible from the cloud-gathering peaks. Our telescope was 

 erected in a small walled enclosure adjoining the house, from which the ground sloped 

 steeply down to the sea in the direction of the sun at eclipse. On the other side it was 

 sheltered by a building. The approximate position was latitude 1 40' N., longitude 

 29m. 32s. E. 



28. The baggage was brought to Sundy on April 28 mainly by tram, but with a break 

 of about a kilometre, where it had to be transported through the wood by native 

 carriers. After a week spent on the preparations, we returned to S. Antonio for the 

 week, May 6-13, as it was undesirable to unpack the mirror so early in the damp 

 climate. On our return to Sundy the installation and adjustments were soon com- 

 pleted, and the first check plates were taken on May 16. Meanwhile the gravana 

 had begun, which, although there is no rain, is generally accompanied by increased 

 cloud. There were, however, some days of clear sky, and the nights were usually 

 clear. 



The coelostat was mounted on a stone pier built for the purpose. The clock weight 

 fell into a pit below the clock deep enough to allow a run of 36 minutes without 

 rewinding. Care was taken to use a particular part of the ccelostat-sector, considered 

 to be the most perfect, in photographing the eclipse and the check field. The telescope 

 (Oxford astrographic object-glass, see p. 295) rested on wooden V's near the two 

 ends, the V's being supported on packing-cases ; the one at the breech-end could be 

 moved laterally to allow of different declination settings, and was marked with an 

 approximate declination scale. A series of exposures of one second was made on a 

 bright star to test whether there was any shake of the telescope after inserting the 

 plate : no shake was detected even when the exposure was made immediately ; but 

 as a safeguard for the eclipse photographs a full second was allowed to elapse before 

 beginning the exposure. The exposure was made by moving a cardboard screen 



