March 29, iqcxj] 



NATURE 



525 



on the birds during the night. Ten of each lot of gnats were 

 dissected after a few days, with the following result : — 



The ten gnats fed on the healthy sparrow contained no 

 "pigmented cells." The ten gnats fed on the sparrow with a 

 moderate number of parasites were found to contain altogether 

 292 " pigmented cells " ; or an average of 29 in each gnat. 

 The ten gnats fed on the sparrow with numerous parasites 

 contained ICXJ9 " pigmented cells" ; or an average of lOO cells 

 in each gnat. These thirty specimens were sent to .Manson in 

 England, who made a similar count of the cells. 



I may mention one more out of several experiments of the 

 same kind. A stock of Cttlex fatigans, all bred from the larva, 

 were fed on the same night partly on two sparrows containing 

 H. relicla, and partly on a crow containing H. daiiiiewskii 

 (placed, of course, under separate mosquito-nettings). Out of 

 23 of the former lot, 22 were found to have pigmented cells ; 

 while out of 16 of the latter, none had them. 



Hence no doubt remained that the "pigmented cells" 

 really constitute a developmental stage in the mosquito of these 

 parasites ; and this view was accepted both by Laveran and 

 Manson, to whom specimens had been sent. In June 1898, 

 Manson published an illustrated paper concerning my re- 

 searches, and showed that the pigmented cells must in fact be 

 the zygotes resulting from the process of fertilisation discovered 

 by MacCullum. 



It remained to follow out the life-history of the zygotes. For 

 this purpose it was immaterial whether I worked with the 

 avian or the human parasites, since these are so extremely 

 like each other. I elected to work with the avian species, 

 chiefly because the plague-scare in Bengal still tendered ob- 

 servations with the human species almost impossible. By feed- 

 ing Citl-'X fatigans on birds with H. relicta and then examining 

 the insects one, two, three and more days afterward-^, it was 

 easy to trace the gradual growth of the zygotes. Their de- 

 velopment briefly is as follows : — After the fertilisation of the 

 macrogamete has taken place in the stomach-cavity of the gnat, 

 the fertilised parasite or zygote has the power of working its 

 way through the mass of blood contained in the stomach, of 

 penetrating the wall of the organ, and of affixing itself q|i, or 

 just under its outer coat. Here it first appears about thirty-six 

 hours after the insect was fed, and is found as a " pigmented 

 cell" — that is, a little oval body, about the size of a large red 

 corpuscle, and containing the granules of melanin possessed by 

 the parent gametocyte from which the macrogamete originally 

 proceeded. In this position it shows no sign of movement, 

 but begins to grow rapidly, to acquire a thickened capsule, and 

 to project from the outer wall of the stomach, to which it is 

 attached, into the body-cavity of the insect host. At the end of 

 six days, if the temperature of the air be sufficiently high (about 

 80° F,), the diameter of the zygote has increased to about eight 

 times what it was a^ first ; that is, to about 60 (jl. If the 

 stomach of an infected insect be extracted at this stage, it can 

 be seen, by a low power of the microscope, to be studded with 

 a number of attached spheres, which have something of the 

 appearance of warts on a finger. These are the large zygotes, 

 which have now reached maturity and which project prominently 

 into the mosquito's body-cavity. 



All this could be ascertained with facility by the method I 

 have mentioned ; and it should be understood that gnats can be 

 kept alive for weeks or even months by feeding them every few 

 days on blood — or, as Bancroft does, on bananas. But a most im- 

 portant point still required study. What happens after the zygotes 

 reach maturity ? I found that each zygote as it increases in size 

 divides into meres, each of which next becomes a blastophore 

 carrying a number of blasts attached to its surface. Finally, the 

 blastophore vanishes, leaving the thick capsule of the zygote 

 packed with thousahds of the blasts. The capsule now ruptures, 

 and allows the blasts to escape into the body-fluids of the 

 insect. 



These blasts, when mature, are seen to be minute filamentous 

 bodies, about 12-16/x in length, of extreme delicacy, and some- 

 what spindle-shaped — that is, tapering at each extremity. Just 

 as the zygotes recall the shape of the Coccidiidce, so do these 

 blasts recall the "falciform bodies." Prof. Herdman and I 

 have adopted this word "blast" for these bodies after careful 

 consideration — but others prefer other names. They are, of 

 course, spores ; but spores which have been produced by a 

 previous sexual process — and are in fact the result of a kind of 

 polyofi bryony. Just as a fertilised ovum gives rise to blasts, which 

 produce the cluster of cells constituting a multicellular animal, so, 



in this case, the fertilised ovum, or. zygote, gives rise to blasts, 

 each of which, however, becomes a separate animal. Prof. 

 Ray Lankester suggests for the blasts of the Hi3emamoebidae the 

 simple term "filiform young." 



At this point the investigations took a turn of extreme interest 

 and importance, scarcely second even to what attached to the 

 first study of the zygotes. Since the blasts are evidently the 

 progeny of the zygotes, they must carry on the life-history of 

 the parasites to a further stage. How do they do so ? What 

 is their function ? Do they escape from the mosquito, and in 

 some manner, director indirect, set up infection in healthy men 

 and birds? Or, if not, what other purpose do they subserve? 

 It was evident that our knowledge of the mode of infection in 

 malarial fever — and perhaps even the prevention of the disease 

 — depended on a reply to these questions. 



As I have said, the zygotes become ripe and rupture about a 

 week after the insect was first infected — scattering the blasts into 

 the body-cavity of the host. What happens next ? It was next 

 seen that by some process, apparently owing to the circulation 

 of the insect's body-fluids (for the blasts themselves appear to be 

 almost without movement), these little bodies find their way 

 into every part of the mosquito — into the juices of its head, 

 thorax, and even legs. Beyond this it was difficult to go. All 

 theory — at least all theory which I felt I could depend upon — 

 had been long left behind, and I could rely only on direct obser- 

 vation. Onat after gnat was sacrificed in the attempt to follow 

 these bodies. At last, while examining the head and thorax of 

 one insect, I found a large gland consisting of a central duct 

 surrounded by large grape like cells. My astonishment was 

 great when I found that many of these cells were closely packed 

 with the blasts (which I may add are not in the least like any 

 normal structures in the mosquito). Now I did not know at 

 that time what this gland is. It was speedily found, however, 

 to be a large racemose gland consisting of six lobes, three 

 lying in each side of the insect's neck. The ducts of the lobes 

 finally unite in a common channel which runs along the under 

 surface of the head and enters the middle stylet, or lancet, of the 

 insect's proboscis. 



It was impossible to avoid the obvious conclusion. Observa- 

 tion after observation always showed that the blasts invariably 

 collect within the cells of this gland. It is the salivary or 

 poison gland of the insect, similar to the salivary gland found in 

 many insects, the function of which, in the gnat, had already 

 been discovered— although I was not aware of the fact. That 

 function is to secrete the fluid which is injected by the insect when 

 it punctures the skin — the fluid which causes the well-known 

 irritation of the puncture, and which is probably meant either to 

 prevent the contraction of the torn capillaries or the coagulation 

 of the ingested blood. The position of the blasts in the cells of 

 this gland could have only one interpretation — wonderful as that 

 interpretation is. The blasts must evidently pass down the ducts 

 of the salivary gland into the wound made by the proboscis of 

 the insect, and thus cause infection in afresh vertebrate host. 



That this actually happens could, fortunately, be proved with- 

 out any difficulty. As I had now been studying the parasites of 

 birds for some months, I possessed a number of birds of different 

 species, the blood of which I had examined from time to time 

 (by pricking the toes with a fine needle). Some of them were 

 infected, and some, of course, were not. Out of III wild 

 sparrows examined by me in Calcutta, I had found H. relicta — 

 the parasite which I had just cultivated in Culex fatigans — in 

 15, or I3"5 per cent. As a rule, non-infected birds were 

 released ; but I generally kept a few to use for the control 

 experiments mentioned above, and the blood of these birds 

 had consequently been examined on several occasions, and 

 had always been found free from parasites. At the end 

 of June I possessed five of these healthy control birds- four 

 sparrows and one weaver-bird. All of them were now 

 carefully examined again and found to be healthy. They 

 were placed in their cages within mosquito-nets, and at the same 

 time a large stock of old infected mosquitoes were released with- 

 in the same nets. By "old infected mosquitoes" I mean 

 mosquitoes which had been previously fed repeatedly on infected 

 birds, and many of which on dissection had been shown to have 

 very large numbers of blasts in their salivary glands. Next 

 morning, numbers of these infected gnats were found gorged 

 with blood, proving that they had indeed bitten the healthy birds 

 during the night. The operation was repeated on several 

 succeeding nights, until each bird had probably been bitten by 

 at least a dozen of the mosquitoes. On July 9, the blood of the 



NO. 1587, \OL. 61] 



