as in life. This worm is the most common parasite of Lake Ellesmere trout, and occurs also in a number of other fishes, some of which belong to widely separated zoological groups.

The following natural hosts have been collected at Lake Ellesmere:—Salmo trutta, Oncorhynchus tschawytscha, Retropinna retropinna, Galaxias attenuatus, Agonostomus forsteri, Rhombosolea retiaria, Rhombosolea plebia, Rhombosolea tapirina, Anguilla aucklandii.

It is usual for worms of this class to have an existence in two different animals, the first of which is taken as food by the host in which the parasite attains maturity, but no intermediate host of Hedruris spinigera is known, and none has been definitely recorded of any species of this genus occurring in other parts of the world. The most consistently infested of the indigenous fishes are the smelt, Retropinna retropinna, and the mullet, Agonostomus forsteri, and if the life history of this worm conforms to that of other closely related forms, it is among the animals forming the food of these fishes that its intermediate host could most profitably be sought. It was found that young smelts of up to 40 mm. in length, taken at Lake Ellesmere, had fed exclusively on copepods and were free from Hedruris, but somewhat larger specimens in which this diet had been supplemented by a few specimens of Tenagomysis novae–zealandiae contained occasional immature worms. The food of adult smelts taken in the lake consisted of Tenagomysis, Paracalliope fluviatilis and small Dipterons, while adult mullets contained Paracalliope, Tenagomysis and gastropods (Potamopyrgus) represented quantitatively in the order mentioned. Of the five food-animals thus associated with the occurrence of the worm, Tenagomysis appears to be the most probable intermediate host, as it is the only one that has been found in all natural hosts except those in which the presence of the parasite may be explained as the result of direct transference from infected food-fishes, and it is, moreover, perfectly satisfactory as regards size. One hundred and seventy of these crustaceans were collected from the lake-bottom and dissected, but all were free from Hedruris, although twenty-four proved to be parasited by trematodes. The remaining food-animals, with the exception of the copepod, which may be definitely ruled out on account of its size, were also subjected to investigation, one hundred specimens of each being dissected, but the only worms discovered were trematodes of somewhat peculiar structure which occurred in several specimens of Potamopyrgus. Several animals that have not been found in the fishes under consideration were also investigated with similar results.

Notwithstanding the failure of these investigations to discover any intermediate host of Hedruris spinigera, there is ample evidence that some preliminary existence is necessary to this worm. The maximum dimensions of the eggs contained in mature females are given by Baylis (1931) as 0.042 mm. by 0.017 mm., and the embryos visible in advanced specimens appear to the present writer to be coiled only once, or, more correctly, doubled. The length of the larva on emergence from the egg can scarcely exceed 0.1 mm., and as the smallest worms found in fishes are approximately 4 mm. in length, there is obviously a considerable gap in the present knowledge

of this parasite's life history. The possibility exists that the larva is free-living during this stage, but no evidence of this has been detected in nettings from the water or in mud and detritus from the lake-bottom.

Table I gives particulars of the food and average length of 66 adult Lake Ellesmere trout classified according to the nature of parasitic infestation. The fish composing the first group were free from parasites of any kind, the second ground contains specimens infested with Hedruris spinigera, and the third those in which the parasite Eustrongylides sp. was present in addition to Hedruris. A consideration of the data presented in this and the following tables suggests that trout become infested with Hedruris only when in the same waters as smelts, and that the parasite is merely transferred to the former fish when the latter is taken as food. There is no material difference in the food of the two parasited groups listed in Table I, each of which contains a considerable proportion of smelts, but it is to be noted that the food of the unaffected group includes only a single specimen of this fish. It is possible that this difference is, to some extent, accidental, and that the food contained by the unaffected group at the time of capture may not be truly representative of the normal diet, but if it is assumed that these trout had recently fed freely on smelts, the absence of Hedruris is difficult to explain. The majority of the adult smelts of Lake Ellesmere are parasited, and in view of the ready adaptability to trout shown by this worm, it is scarcely conceivable that smelts could be taken, even in moderate quantities, without some degree of infestation resulting. It seems more probable that certain trout, either from choice or owing to peculiarities of individual habitat, adopt feeding habits differing from those of the majority, and that freedom from infestation is maintained so long as the natural hosts of Hedruris are excluded from the diet.

The 34 immature fish detailed in Table 2 were taken in early summer, about the upper limit of the deep section of the Selwyn, and appear never to have been lower than this point. The scales of these fish show perfectly open edges, which indicates that growth was proceeding freely at the time of capture, and the uniformity of the structure formed since the winter supports the view that the same locality had been inhabited during the last five or six months. The presence of Hedruris in ten of these fish is quite consistent with the explanation that the parasite is obtained from smelts taken as food, as the summer range of the latter fish extends to the locality under discussion, and specimens taken at this point were found to be almost universally infested. The difference between the food of these trout and that detailed in Table I appears to be primarily one of environment. At the point where the present specimens were taken the country is for the greater part grassed and heavily infested with brown beetles, Odontria zealandica, conditions that do not exist on the lake-flats near the mouth of the river. During summer evenings these beetles resort to the willow trees bordering the stream at this point, and many are precipitated into the water, where they are freely taken by trout. Considerable numbers of caddis larvae are also taken from the nearby shingle reaches.