DELPHINIUM SCOPULORUM. 23 



same trip. B. juxtus sometimes visits withered flowers, stopping to seek 

 nectar at those in which the lower side petals have dropped off. B. edwardsi 

 and morrisoni land and take nectar in the same manner as juxtus. As 

 Thanaos martialis lands on the lower sepals, the delicate ligule is pushed 

 into the nectary and the hairs on the under side of the head come in contact 

 with the anthers. 



Robertson (1889:120) states that the white upper petals are a sure guide 

 to the nectar in Delphinium tricorne, in which the lateral petals serve to 

 protect the pollen. It agrees with D. datum and D. consolida in being 

 protogynous and especially adapted to bumble-bees. In Europe these 

 species are exclusive bumble-bee flowers (Knuth, 1908:44), but Robertson 

 notes several other genera of bees and a number of butterflies and moths. 



EXPERIMENTS. 



CHANGE OF POSITION. 



Inverted racemes. — Bombus juxtus hovered above 6 flowers which 

 were inverted, but apparently noticed that something was wrong and 

 did not land. At the seventh flower it made an attempt at landing, but 

 found the pointed front part of the spur too narrow. After slipping off 

 twice when it tried to put its hindlegs in this spur, it flew away. Other 

 individuals flew above these flowers, but did not alight. 



Horizontal racemes. — In these flowers the hood pointed down and 

 Bo?nbus juxtus was not bothered by the change. It landed and found the 

 nectar easily, using a side sepal and one of the lower ones for a landing 

 platform. 



MUTILATION. 



Cotton plugs. — Absorbent cotton was placed in the opening to the spur, 

 which serves for the nectary. Bombus juxtus stopped, but when it found 

 no opening did not attempt to probe around the cotton, but went to the 

 next flower. A slender cotton roll was placed in the nectary in such a way 

 that it projected 1 cm. in front of the flower and changed the whole aspect 

 of the latter. This bee landed on the cotton projection and tried to find 

 the opening to the nectary, but failed. It repeated the performance on 

 the next flower with the same results. 



Petals removed. — The petals, which recurve and protect the stamens, 

 were removed, thus making the black group of anthers very conspicuous. 

 The curved white filaments showed above the anthers, and the nectary 

 openings were visible above the filaments, thus changing the aspect of the 

 flower materially. On the first day of this experiment, Bombus juxtus 

 took no notice of the change and found the nectary without difficulty. 

 On another day the results were quite the reverse and 5 individuals passed 

 by these flowers during a period of 15 minutes. B. edwardsi went to 7 of 

 the mutilated flowers, apparently without noticing the difference (plate 16). 



Spur removed. — Bombus juxtus pushed its tongue back and forth 9 

 times in one of these flowers, the tongue projecting 6 to 7 mm. beyond the 

 cut end. In another instance, the tongue was rapidly pushed back and 

 forth 4 times through the opening. A third individual noticed the change 

 in the flowers and did not land. 



