8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



The preceding preliminary experiments clearly show the following : 

 ( 1) In regard to foods bees have likes and dislikes ; (2) before show- 

 ing preferences between foods bees always eat more or less of them 

 first, unless the foods contain strong repellents; (3) the longer the 

 bees are confined in the experimental cases the less they eat, and (4) 

 some of the substances fed are injurious to them. For the last two 

 reasons only the first five counts are sufficiently reliable for determin- 

 ing the total average per cent of bees eating at any one count. These 

 experiments indicate that bees may have a sense of taste, because 

 neither the direction of the light nor the arrangement of the food in 

 the cases helps in selecting the food they like best, and the olfactory 

 sense may not be the sole factor in selecting foods, for bees must 

 usually eat more or less of them before being able to show preferences 

 between them. It is probable that bees cease eating some foods 

 because their alimentary tracts may be affected, and for this reason 

 alone they may reject the particular food that does not agree with 

 them. 



The preceding results suggest five classes of foods to be used in the 

 following experiments. Foods containing strong repellents may be 

 employed to determine the importance of the olfactory sense in caus- 

 ing bees to avoid such substances, and foods containing sweet, bitter, 

 sour, and salty substances may be used to ascertain if bees show pref- 

 erences between foods having the four attributes of human taste. 



2, Experiments in Feeding Bees Foods Containing Repellents 



Pure cane-sugar and candy containing oil of peppermint (de- 

 scribed above, page 4) were fed to fresh bees in the cases as 

 described in the preceding pages. After waiting 10 minutes the first 

 count was recorded, and thereafter every 30 minutes. As an average 

 for the 100 bees for five counts, 35.4 per cent were seen eating the 

 pure cane-sugar candy at any one count, while they never touched the 

 candy containing oil of peppermint. 



The preceding was repeated by feeding candy containing carbolic 

 acid (described on p. 6) and pure cane-sugar candy to fresh bees. 

 As an average for the 100 bees for five counts, 41.4 per cent were seen 

 eating pure cane-sugar at any one count, while none touched the 

 candy containing carbolic acid. 



The preceding was repeated by feeding pure honey and honey 

 containing whiskey to fresh bees. Four grams of pure honey were 



