RECENT INVESTIGATIONS. 199 



and 20 for the normal control. When a single petal was placed on the de- 

 corollate flower, the latter yielded 8 inspections to 9 for a single petal 

 pinned to the stem and 27 for a normal flower; changing the relative 

 position of the three gave 27, 8, and 22 inspections. Placing the smaller 

 brighter yellow blossom of an Asiatic cotton between two normal flowers 

 of the American variety resulted in 14 inspections for the former and 20 

 and 29 for the latter in one case, and 10, 16, and 20 in another, while putting 

 the foreign flower at one end of the series gave 3, 10, and 11 respectively 

 in one instance and 11, 12, 9 in the other. A day-old blossom of a deep 

 reddish-purple color received 12 visits to 26 for a recently opened cream- 

 colored one. Four experiments with two controls and a normal blossom 

 with honey added at the base of the petals, arranged in varying order, 

 gave respectively 62, 64, and 66 inspections, the actual visits being but 2 

 for the flower with honey to 27 for the other two. 



The petals of a normal bloom were covered on both sides with cotton 

 leaves cut to match and a drop or two of honey added. When this was 

 exposed with a control with honey and one without, the inspections were 

 respectively 0, 34, and 25, the actual entrances being 0, 2, and 6. In the 

 next installation the pieces of leaf were removed from the outside of the 

 corolla and this was exposed with a normal flower and one with the petals 

 removed. The respective inspections were 9, 11, and 1, while the actual 

 visits were negligible. When one bud due to open the next day was pulled 

 fully open and another but partly so, the first was inspected 15 times to 

 twice for the latter and 24 for the normal, the latter receiving 14 visits 

 to none for the other two. To entirely eliminate the effect of odor, a glass 

 plate was fixed before a flower and leaves drawn in about the edges to con- 

 ceal most of the plate but leave the flower fully visible. This received 18 

 inspections in contrast to 37 for the normal. When a flower was placed 

 in a box with a glass face, it was inspected twice to 5 times for the control. 

 With a crepe-paper blossom of the proper color added, the boxed flower 

 was inspected once, the normal 8 times, and the imitation 3 times. In the 

 final experiments, detached petals were usualty inspected, but the numbers 

 were too small to be very significant. 



The simplicity and variety of Allard's experiments, coupled with the 

 use of single flowers and constant controls, give his results an excep- 

 tional value. They emphasize the attractive value of the color at the ex- 

 pense of odor and suggest that the latter merely guides bees from a distance 

 or from one field to another. The discrimination between artificial flowers 

 and normal ones is thought to depend upon perceptible differences in color 

 and texture rather than to the presence of repellent odors from the material 

 and support the view that bees develop keen powers of discernment as 

 to the structural details of flowers. 



Color sense and memory in the honey-bee. — Dobkiewicz (1912:664) 

 carried out three series of experiments dealing respectively with the color- 

 • sense, memory, and learning ability of the honey-bee. The first series 

 employed crude artificial flowers of yellow or red paper placed at the same 

 spot in a clover field. No visits were made to installations of 10 yellow or 

 10 red artefacts without honey, or at first to 5 of each color with and with- 

 out honey. For the next experiment the blue artefacts were removed, 



