34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 82 



direct sunshine. As suspected under these conditions, the number of 

 beetles counted in the four positions varied greatly. More beetles 

 were always counted in the outer positions (Nos. i and 4) than in 

 the inner ones (Nos. 3 and 4). This was caused largely by the beetles 

 following the sides of the cage while moving toward the window. 

 In 40 series of tests, selected at random, 67.04 per cent of the beetles 

 were counted in the outer positions and 32.96 per cent in the inner 

 positions. The following percentages were counted in the four posi- 

 tions: 37.06 per cent in position i ; 17.01 per cent in position 2; 15.95 

 per cent in position 3 ; and 29.98 per cent in position 4. Since each 

 substance used lay in all four positions during any one series of tests, 

 these large differences did not supposedly change the arithmetic mean, 

 but they greatly affected the probable error, because each number of 

 beetles counted on a substance was considered a statistical item. Since 

 the beetles ate more freely during the forenoon than during the after- 

 noon, the probable errors were further affected. 



In addition to the preceding statements, the general plan in con- 

 ducting these tests was to put 60 beetles of approximately the same 

 age and physiological condition in each cage. The number of beetles 

 on (or touching) the food was counted at intervals of 45 minutes, 

 and this number was considered a statistical item. The food was re- 

 newed whenever necessary to keep it in an appetizing condition, and 

 to prevent contamination it was usually put on unused pieces of card- 

 board. Since the beetles had a tendency to congregate at the ends of 

 the cages nearest the window, the cages were often turned end for 

 end, thus causing the insects to scatter more evenly. The daily tem- 

 perature and relative humidity in the laboratory were recorded, and 

 a record of the outside climatic conditions was also kept. In brief, 

 everything possible was done to obtain reliable data which could be 

 treated statistically. The arithmetic mean and probable error are 

 stated in tables 2 to 11 for reference in connection with the following 

 discussion. Since the statistical items were never less than 16, the 

 following formula for calculating the probable error was used — • 



P. E. m= ±0.6745- 



VN 



(i) Beetles can distinguish differences between water and salty 

 liquids. — To determine whether Mexican bean beetles " like " or " dis- 

 like " the four classes of substances which produce the four human 

 attributes of taste, many series of tests were conducted. The results 

 obtained are given in tables 2 to 5. 



To ascertain whether these insects " like " salty water, sodium 

 chloride, potassium nitrate, and magnesium sulphate (epsom salts) 



