CELASTEINA AEGIOLUS. 447 



the ants are away from the last segments, and are retracted when 

 they come Dear. I counted the length of these periods of complete 

 and quiet expansion, 10, 20, 50, and to 80 seconds, the period always 

 ending with the approach of the ants. I experimented, .... 

 placing larvae .... upon stems of the growing plants, where the ants 



had access to them As soon as the ants discovered one of them, 



there was an immense excitement, and a rush for the last larval 

 segments. The larva forthwith relieved itself by the execretion of the 

 fluid, and the tubes stood out with tops expanded during the periods. 

 If I placed a fresh larva on a stem on which were no ants, there 

 was no excitement in the larva, no appearance of the tubes, and 

 no movement in the median gland. If ants were now transferred 

 to the stem, at once the larva changed its behaviour. The ants were 

 only noticed as attending the larvae in the later stages of the latter, 

 and only observed in the case of summer-feeding larvae. Edwards 

 further notes (op. cit., p. 13) that the ants, when confined with 

 larvae in the 1st instar, treated them with indifference. He intro- 

 duced ants to larvae in separate glass-tubes, some larvae being at the 

 middle, and some near the end of the second stage (i.e., near the second 

 moult); one of the larvae was caressed several times but no tube appeared : 

 another larva objected to the ant, thrashed its anterior segments about, 

 and the ant left it. An ant introduced a day after the third moult . . . 

 solicited as usual ; the tubes appeared, and a drop of fluid came 

 from the honey-gland, which the ant drank eagerly; it returned several 

 times but obtained no more ; on the same day an ant was introduced to 

 two larvae towards end of third stage (i.e., just before third moult), 

 there was a slight movement of the tubes in one larva, a mere point 

 protruding, but no fluid from the honey-gland. The other larva did not 

 respond at all, and the ant left both. An ant was introduced with 

 another larva in its third stage, when the tubes were seen to play actively, 

 but, though the ant held its mouth to the honey-gland for some seconds, 

 no excretion was observed. It is in the last larval stage that the fluid 

 flows freely at the solicitations of the ants. Edwards insists, however, 

 that this only occurs with summer larvae, on Cimicifnga raeemosa, the 

 flower of which is exceedingly sweet ; he has not observed an ant on 

 dogwood, and ants placed with larvae feeding on dogwood soon became 

 indifferent to them ; similarly, ants -attracted to autumn larvae, feeding 

 on Actinomeris, which has a flower bitter to the taste, have been noticed 

 to turn away after investigation. Edwards considers that the tubes 

 (on the 8th abdominal segment) serve as attraction signals to the 

 honey-gland on the 7th abdominal. He has observed larvae in the last 

 stage, w T hen no ants have been present, irregularly protrude the tubes, 

 without any corresponding activity of the honey-gland ; the presence of 

 the ants, he says, seems necessary to produce this and the larvae appeared 

 only to emit the fluid when the ants were near. McCook informed Edwards 

 that, in the spring of 1877, he saw a small green larva on Cimicifuga 

 raeemosa, and a black ant attending it, stroking the tail incessantly, 

 moving away, and returning to go through the same process ; he 

 watched this for two hours, and saw that the purpose of the ant was 

 at least friendly, although he was at a loss to explain these strange 

 manipulations. 



Foodplants. — Flowerbuds, young green berries and tender young 

 leaves of Ilex aquifolium , flower-buds and young leaves of Heeler a helix, 



