62 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO5 



to the eastward, but later an effort was made to control possible altera- 

 tion of the exposure due to their growth. In more recent years this 

 policy had to be abandoned, and trees were gradually removed because 

 some became so high that trimming would no longer suffice to prevent 

 sheltering of the sunshine recorder and rain gages, while others were 

 an obstruction to the erection of new buildings. Whereas the palms 

 partly shaded the screen for many hours of the day during the first 

 10 or 20 years, those that are left now are some distance to the south 

 and provide no protection from the sun at all. The distance from the 

 coast has altered also, the shore line having built out with the gradual 

 deposition of sand, until now the screen is 50 ft. from it. Another 

 factor which may have had a slight influence on the exposure is the 

 increase in the number of wooden buildings 60 ft. to 80 ft. to the east 

 of the site. 



The original screen, photographs of which were printed in another 

 publication, was an "English " one built in Gottingen and brought to 

 Samoa by Dr. Otto Tetens. The inner space of this screen was 3 ft. 

 wide, 2 ft. deep, and a little over 3 ft. high. Double louvres were 

 fitted on the eastern and western sides, but the doors on the front and 

 back had only single ones until 1905, when Dr. Linke had them fitted 

 with a second inner set. The roof was double, the outer one sloping 

 up to the south at an angle of about 15°. Air slits in the bottom facili- 

 tated circulation inside the screen, which was mounted on wooden 

 posts 6 ft. high. 



The readings of the thermometers in this screen, which had no 

 auxiliary protection, were compared with those of an Assmann psy- 

 chrometer at 7 a. m., 2 p. m., and 9 p. m., at the time of the year when 

 the most adverse conditions for the recording of true temperatures 

 were thought to obtain. The mean corrections to the screen readings 

 were 0.00, —0.26, and +0.02° C, respectively. 



Sometime after 1908, but before 1920, the screen just described was 

 replaced by one of a different type. The design of the latter, in which 

 the standard set of thermometers are still exposed, was evidently con- 

 sidered to be more suitable for the tropical climate. It consists of a 

 louvred box or cupboard, the clear internal dimensions of which are : 

 height, 20 in.; width, 20^ in.; and depth, 17 J in. 



Since August 1932 a second set of thermometers have been exposed 

 in a Stevenson screen of approved pattern. 



Reference to the mid-temperature, defined as i(Max.-rMin.). 

 shows that, on the whole, temperatures recorded in the Stevenson 

 screen are slightly higher than those in the tropical. 



