12 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



levulose, none eating dextrose and 1 .8 per cent eating raffinose candy, 

 making a total average of 21.8 per cent eating at any one count. 



Dextrose, raffinose and dextrine candies were used in the same way. 

 As an average for the 100 bees for five counts, 21 per cent were seen 

 eating raffinose, 12 per cent eating dextrose and 6.8 per cent eating 

 dextrine candy, making a total average of 39.8 per cent eating at any 

 one count. 



Dextrine, lactose and maltose candies were used in the same way. 

 As an average for the 100 bees for five counts, 42 per cent were 

 observed eating maltose candy at any one count, but none was seen 

 eating lactose or dextrine candy. 



The preferences shown between these candies may have been par- 

 tially due to the amount of water in them. No two of these candies 

 absorbed the same amount of water vapor from the air, but during 

 the first day the water in any of them was not noticeable, although 

 after that it was quite noticeable. Levulose absorbed the most water 

 vapor and saccharine the least. 



Dextrose, raffinose and maltose candies were next used. As an 

 average for the 100 bees for five counts, 16 per cent were seen eating 

 maltose, 12 per cent eating raffinose and 7 per cent eating dextrose 

 candy, making a total average of 35 per cent eating at any one count. 



To ascertain if bees show preferences between honeys, an equal 

 amount of light-colored honey and dark-colored honey was poured 

 into each of five feeders. Perhaps most of the light-colored honey 

 came from basswood trees, while the source of the dark-colored honey 

 was unknown. The latter honey was taken in the crystallized form 

 from old combs and was then melted. The odors and tastes of these 

 two honeys were quite different. Fresh bees from the alighting- 

 boards were introduced into the cases, and during the first five minutes 

 after giving them the two honeys, they ate each one greedily. By the 

 time they had eaten five minutes, most of them had selected the honey 

 they liked the better. At this stage the ones eating were counted, 

 and five minutes later were counted again. After this few were seen 

 eating, because nearly all of them by this time had filled their honey 

 stomachs. This set of experiments was repeated twice. As an aver- 

 age for the 300 bees for two counts, 24.3 per cent were seen eating the 

 light-colored honey and 18,8 per cent the dark-colored honey, making 

 a total average of 43.1 per cent eating at any one count. 



Fresh bees were placed in the cases, and they were fed light- 

 colored honey and sugar syrup (half sugar and half water) in the 

 same manner as just described. As an average for the 100 bees for 



