parrut}] .THE MECHANISMS OF FLOWERS. 259 
the last. Thus all the three kinds of stigmas will tend in time to 
"receive pollen from stamens of their own length. If the middle- 
sized and longest organs were also inclosed within the tube, the 
| proboscis or head of the insect in passing down would come in 
contact with all, one after the other, and cross-fertilisation could not 
"be e nearly so well restricted to its proper lines. In point of fact, the 
insect alighting on the middle-sized and longest organs never has 
_ occasion to touch these with its proboscis or the fore-part of its 
ventral surface, the parts that are to come in contact with the 
shortest organs; nor is the posterior part of the ventral surface 
which has been in contact with the longest organs, ever brought 
forward during the process of sucking to touch the intermediate 
ones. So that, as a rule, “ legitimate crossing” only is performed. 
In the action just described, which leads regularly to legitimate 
© ossing, only the larger and middle-sized bees and the eee flies 
take part. 
_ I have observed : (1) Cilissa melanura, Nyl. ¢ and 9, wherever Lythrum 
Salicaria grows, both sucking honey and gathering pollen, and almost confining 
‘itself to this one plant. (Since its proboscis is only 3—4 mm. long, it must thrust 
Beret part of its head, which is 2—3 mm. broad, into the tube : it then touches 
the shortest reproductive organs with the under surface of its head, the next 
_ with the ventral surface of its thorax, and the longest with the ventral surface 
_ of its abdomen ; so that its dimensions suit the flower, just as well as the flower 
' must suit it, for the insect to confine its visits so exclusively to it. I have only 
found one exception to this exclusiveness, for I once found a male of Cilissa 
melanura sucking honey on Thrincia hirta, D.C.) ; (2) Saropoda rotundata, Pz. 
; 3) frequently (body without proboscis 10—1I mm., proboscis 9 mm. long) ; 
3) Apis mellifica, L. § (body without proboscis 1I—13 mm., proboscis 6 mm. 
long) ; (4) Bombus terrestris, L. § (body 12—16 mm. S seoboseis 7—9 mm. ); (5) 
'B. agrorum, F. § (body 10—13 mm., proboscis 9—11 mm.) ; (6) B. silvarum, 
. § (body 10—12 mm., proboscis 10—12 mm.). All the humble-bees were 
1 equent, and, like Nos. 2 and 3, sucking honey only ; (7) Megachile centun- 
ew aris, L. ¢ (body 10 mm., proboscis 6—7 mm.), also sucking honey. The 
ast six species, as the aimexed measurements show, have all. tongues long 
| enough to reach the honey by thrusting in the tongue alone ; they all therefore 
touched the shortest organs with their tongues, the next series with the lower 
' Surface of the head, and the longest with a part of the ventral surface of the 
| body 4—5 mm. further back. The hive-bee must, in full-sized flowers, have 
arust the extremity of its head 1 mm. deep into the calyx-tube. 
__ Besides these bees, which deserve to be placed in the first rank as legitimate 
fertilisers, the Loosestrife is visited by a number of long-tongued flies, especially 
| Rhingia rostrata. This fly, standing on one or more of the petals, after gently 
rubbing its forefeet together, and brushing its tongue and head with them, 
stretches its proboscis out to a length of 11—12 mm., and thrusts it down into 
the flower, letting it remain there from six to ten sheonls Immediately after 
withdrawing it from the tube, it usually manipulates one of the anthers with its 
Ss 2 
