198 PRINCIPLES AND CONCLUSIONS. 



Pattern vision in the honey-bee. — In experiments on this subject, 

 Turner (1911:249)— 



"Made use of pasteboard boxes like those employed in his experiments on color 

 vision in the same insect. He constructed artefacts showing seven different color 

 patterns. Bees which had been trained to gather honey from one pattern were tested 

 to see if they could choose this pattern from one or more of the others. In some cases 

 the artefact of the pattern to be chosen contained honey, while the rest had none; 

 in other cases there was honey on some of the artefacts of all patterns, and in still 

 others there was no honey on any of the artefacts. Five hundred and eight correct 

 selections out of 518 were made, indicating that color patterns are perceived by bees. 

 Since they can distinguish both color and pattern, no evidence can be drawn from the 

 visual powers of bees against the hypothesis that colors and patterns in flowers are 

 adapted to secure the visits of insects." 



Experiments with cotton blossoms. — Allard (1911:607) has made 

 a large number of interesting experiments with single cotton blossoms. 

 The normal flower, modified flower, and control were disposed at the corners 

 of a triangle and about 4 feet apart, or in a line in the same row. Petals 

 pinned to a stem received as many inspections as the normal flower, almost 

 all of them being made by Melissodes. The number of entrances was less 

 than one-fifteenth the number of inspections, and at all times the actual 

 visits were few. When the petals of one of the same group of flowers were 

 removed, the inspections dropped from 81 to 4, though they actually in- 

 creased for the other normal flower. Replacing the petals brought the 

 inspections up to those for the normal, but when cloth petals from an artificial 

 rose were employed, the number dropped to 4 in comparison with 48 for the 

 normal control and 65 for the petals alone. Five cotton petals were then 

 placed over the cloth ones and the number of inspections again rose to 

 a point slightly above that for the control. The next installation con- 

 sisted of a normal blossom in situ, one pinned in position, and an 

 artificial half-opened bud made by pinning normal petals together and 

 wrapping a piece of green cotton leaf about the base. The latter received 

 practically the same number of inspections as the most visible control. This 

 was next modified by concealing one of the normal flowers so that it was 

 visible only from above; this yielded 1 inspection to 12 for the control 

 and 40 for the artificial bud. Removing the disguise caused the flower 

 to be inspected twice as often as the control. The substitution of crepe- 

 paper petals led to the reduction of inspections to 2 in contrast to 16 for 

 the control, while placing three real petals on the paper ones increased 

 them to 11 in comparison to 7 for the control and 21 for the artificial bud. 



The following installation comprised a single real petal pinned to a 

 stem, an artificial bud made by wrapping a portion of a cotton leaf about 

 the base of 5 petals rolled together and a normal in situ; these received 

 2, 8, and 3 inspections respectively. A single petal in competition with 

 a normal blossom alone obtained 16 inspections to 26 for the latter. In 

 the following experiments the flowers were on 3 consecutive plants in the 

 same row. The first consisted of a flower of bindweed between two normal 

 flowers pinned in position, which attracted attention as often as the nor- 

 mals, but was entered less frequently. When the petals were removed 

 from one of the controls, this received no inspections to 19 for the bindweed 



