ROBERTSON FLOWERS * INSECTS UMBEI.LIFER^. 455 



Cicuta maciilata, L.- — I'ive umbels of each order bore an average of 

 flowers and umbel lets as follows : 



1st order — 4 male and 4S5 hermaphrodite flowers in 16 umbellets. 

 2d order — 127 " " 824 " " '' 28 '' 



3d order — 360 " " 448 " " " 27 " 



4th order — 438 " " 43 " " " 22 " 



As a rule, umbels of ist order bear only hermaphrodite flowers, those 

 of 4th order only male flowers. The hermaphrodite flowers are proter- 

 androus. 



This plant bears many large umbels of white flowers with fully exposed 

 nectar. Cicuta was observed nearly three times as much as Ileracleum, 

 but Ileracleum shows 15 more flies. As we have observed, Heracleum has 

 20 more flies than hymenoptera, but Cicuta has 71 more hymenoptera than 

 flies, nearly twice as many. While diptera reach their maximum on Ile- 

 racleum, hymenoptera reach their maximum on Cicuta. It is the first to 

 show BembecidJE. Compared with Pastinaca, which blooms earlier, it 

 shows an increase in all hymenoptera except Parasitica, Philanthidae, 

 Ermenidie, Andrenidae, and Apidje. The Parasitica would no doubtequal 

 the numbers taken on Pastinica, or Slum, if I had given the sam^ atten- 

 tion to the collection of them, and then the extent of the preponderance 

 of hymenoptera on Cicuta would have been better indicated. 



Eryngiutn yucc<sfolium, Mex. — Agrees in general with E. campestre, 

 as described and figured by Miiller in the "Fertilization of Flowers," 271. 

 Although the styles are strongly exserted from the first, they do not seem 

 to become receptive until after the pollen is discharged, so that the flower 

 is proterandrous. The early elongation of the styles seems to be to ob- 

 struct the passage to the nectar, so as more effectually to exclude short- 

 tongued insects. The nectar is more deeply seated than in any other 

 species here considered, and, as a consequence, we find a marked increase 

 in insects of large size and long tongues. Compared with Zizia, we find 

 a smaller number and proportion of bees, on account of the increased 

 abundance of other hymenoptera. We also find fewer diptera, on account 

 of deeper honey and competition of hymenoptera." 



But the effect of deep-seated honey is best shown by comparison with 

 Cicuta, which was observed nearly equally and under equally favorable 

 conditions, and, since they bloom at the same time, they are exposed to 

 the same kinds of insects, so that the differences in the lists may be re- 

 ferred to differences in floral structure. The Parasitica are limited to 

 Leucospis. Chrysididse and Nyssonidse are absent. Pompihdse fall from 

 20 to 6; Larridse from 12 to 3; Crabronidce from 14 to 4; Sphecida; from 12 

 to 8. The larger and longer-tongued species — Scoliidje, Philanthidae, and 

 Eumenidse — remain about the same, while the large Bembecidje show an 

 increase. There is marked increase in long-tongues, such as Apida?, Bom- 

 bylidiE, Conopida-, and lepidoptera. Indeed, the list shows more Bom- 



