C. H. TURNIvR. 



paragraph of Series I.] After a bee had made several trips to 

 tin- plot, it would rly directly to it without making the compli- 

 cated searching movements. This striking behavior furnished a 

 sure- means ol telling when a new bee had formed the association. 

 I changed the arrangement of the individual artifacts so often 

 th.it tlu- bees usually were forced to search for the honey-produc- 

 ing box. After tin- asocial ion had been well established, the 

 bees ii-ually departed tor home without making a careful flight 

 of orientation. It,ho\\c\er, I had made a marked change in the 

 position of the artitact since the last visit of the bee, then a care- 

 ful Might of orientation was always made. Xo one who had 

 watched these pronounced searching movements could doubt that 

 bee- are largely guided by memory pictures of the environment. 



After a bee had learned, by experience, that artifacts bearing a 

 certain color pattern contained a more copious supply of easily 

 obtained honey than ordinary flowers, 'it would select artifacts 

 bearing that color pattern from those marked in a different way. 

 This was true: (i) when several of the artifacts to be selected 

 \\ere scattered among a number of plain artifacts of the color- 

 used in making the color-pattern [Ex. 6-1 1] ; (2) when the artifact 

 to be selected was scattered among several other artifacts, some of 

 which were plain and some of which were marked with patterns 

 unlike that of the artifact to be chosen [Ex. 12-15, 17, 1 8]; (3) when 

 tin- only difference between the artifacts was that one was marked 

 with transverse and the other with longitudinal stripes [Ex. id): 

 i when the artifact to be selected contained the honey and 

 the others did not [Ex. 8, 12-17]; (5) when honey was to be found 

 not only in the artifact to be selected, but in some of the other 

 artifacts also [Ex. 9]; (6) when none of the artifacts contained 

 honey |Ex. to, II, 18, 19]. 



All of the facts mentioned in the above paragraph are recorded 

 in the follou in- table. 



Although this table shows conclusively that bees can recognize 

 color patterns it does not tell the whole story. It does not show 

 that t\\o of tin- mi-takes \\ere made in selecting a box mottled 

 with red and green for one marked with longitudinal red and green 

 stripes; that two mistakes cons'sted in selecting a box marked 

 with black and \\hite longitudinal stripes lor one marked with 



