Investigation of the Stomach Contents of New Zealand Fresh-water Shags.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 74, 1944-45, Page 320

Investigation of the Stomach Contents of New Zealand Fresh-water Shags. By R. A. Falla and G. Stokell. [Read before the Canterbury Branch, July 5th, 1944; received by the Editor, September 9, 1944; issued separately, March, 1945.] Introduction. The Cormorants of New Zealand, or Shags, as they are more commonly called, can be divided into three groups for which a convenient distinguishing key character is the colour of the feet. 1. The yellow-footed group, consisting of the closely related species Phalacrocorax punctatus, P. steadi, and P. feather-stoni. For these the sub-generic Stictocarbo may be used. They have a restricted distribution on rocky coasts adjacent to deep water, and the last two have a narrow geographical range at Stewart Island and the Chatham Islands respectively. 2. The Flesh-footed or Pink-footed group. These are the sub-Antarctic shags, with a local race inhabiting each of several islands. They are exclusively marine and their affinities with several circum-polar species are fairly close. The New Zealand forms, however, are sufficiently distinct to be considered endemic. The species at present recognised are:— P. carunculatus—Western Cook Strait. P. chalconotus (including P. huttoni)—Otago and Stewart Island. P. ranfurlyi—Bounty Islands. P. onslowi—Chatham Islands. P. colensoi—Auckland Islands. P. campbelli—Campbell Islands. 3. Black-footed Shags. This group consists of the cosmopolitan species P. carbo (the Black Shag), P. varius (the Pied Shag). P. brevirostris (the White-throated or Little Pied Shag) and P. sulcirostris (the Little Black Shag). These species all occur in Australia, and the New Zealand races are hardly distinguishable. None of them has an exclusively fresh-water range but all inhabit the lake and river systems in varying degrees. The Pied Shag does not as a rule range beyond estuaries. The Little Black Shag has a restricted fresh-water range in the North Island only, while the other two range inland throughout New Zealand, and the Black Shag in particular may be found on any stream or lake. Although thus widely distributed, however, it cannot at the present time be considered abundant. Distribution of Black Shags. It is evident from general summaries, such as that of Stead (1932), that the Black Shag is much less abundant in New Zealand than formerly. It is difficult to get comprehensive information on the number of breeding colonies but this could probably be done in the near future owing to the Dominion wide coverage of the recently

formed Ornithological Society of New Zealand. From the publications of the first three years of activity of this Society, 1941–44, the following information has been derived. From North Auckland there are no records of large numbers, and only some colonies of less than a dozen nests are recorded by observers in this area. In the Lower Waikato the nesting headquarters appear to be Lake Waikare, where the number of nests, perhaps 200 in 1934, appears not to have increased in the last 10 years. From Taranaki and Northern Hawke's Bay no recent records are available, but for Southern Hawke's Bay Mr C. A. Fleming has contributed the following (Ann. Rept. O.S.N.Z., 1941): “In six months in S. Hawke's Bay every Black Shag seen was noted. Though anglers report this species as abundant in the Manawatu and other trout streams, there were only 42 birds recorded in-this area, and less than 100 in the whole district. Odd birds were seen at Kumeroa, Mangatiwainui, Mangatainoka, Dannevirke, Mangapuaka, Waustead, Purimu Lake, Blackhead (coast) and lakes behind Blackhead. Flocks (over 12) at Manawatu near Tamaki Junction and north of Kumeroa, where there are regular roosting-places, and at Poanui, N. of Pourerere, where, on 23/5/41, numbers of birds had fine white flank patches and were making regular trips with nesting-material from Poanui to a point at least two miles north, where undoubtedly there is a nesting-colony. A further colony is reported at Hatuma Lake, south of Waipukurau.” From Wairarapa and Palliser districts small colonies of up to 30 nests that have not increased over a number of years are reported, but at Gollans Valley, near Wellington, a colony observed at intervals by one of us (R. A. F.) since 1930, when there were 40 occupied nests, had dwindled to 20 nests in 1932, 12 in 1934, and finally had been found abandoned in 1942. It is not known if any survivors moved elsewhere, but the indications are that this colony has been exterminated. The numbers recorded in Nelson and Marlborough are somewhat larger than from other districts, and flights of up to 300 birds have been recorded moving from the Waimea Plains towards the coastal region of the Sounds. The resident population of Marlborough would also number several hundreds. In Canterbury the coastal cliffs east of Lake Ellesmere and Forsyth still afford a nesting sanctuary for a scattered colony of several hundred birds. In the back country, nesting colonies are widely separated and of comparatively small size, such as those of the Wilberforce River and at Waimate, from which samples have been examined in the preparation of this paper. No full records are available from Westland, Otago, and Southland, but the species is fairly common in all three districts. Method of Presenting Results. Adequate investigation of the feeding habits of cormorants in the lake and river systems of New Zealand is long overdue. In comparison with the vigorous expression in the popular press of a prejudice against shags that has now become traditional, there is to be found as yet no publication of a detailed analysis of shag food

in New Zealand that could be used with confidence by a student of ornithology or of the conservation of a balanced fresh water fauna. The present contribution does not claim to make good the deficiency: the data offered, although significant, are both meagre and incomplete. In a paper read in 1937, but published only in abstract, one of us (Falla, 1937) has outlined the requirements of such an investigation as (a) A census of the nesting colonies of each species. (b) Determination of the complete annual food cycle of each species. (c) Analysis of results to determine whether in the ecology of streams and lakes any cormorant food is taken to such excess as to cause depletion. At the time of writing, none of these conditions has been fulfilled, but some significant samples have been analysed from four separate districts, and it is considered that a listing of these results gives enough information about the feeding habits of shags on inland waterways to warrant publication. The available records are here presented separately for each locality investigated. It seems that they would in this way have some comparative value, as the conditions of food supply in various areas are quite dissimilar. The numerical method of recording stomach contents has been followed as it seems on the whole better than any attempt to make volumetric analysis or to indicate the relative bulk of the various foods. The individual food organism is therefore regarded as the unit. Where available, some details of the exact locality and time of day at which the birds were collected are given. On this point, Dr. Scaife Armstrong has remarked in a letter covering the forwarding of his results: “I have put in the locality where the birds were shot as it seems to me that this is of first importance in the type of food taken, for the birds go to definite fishing grounds for certain kinds of food and then when fully fed go to a resting place where they are shot. If the birds are shot on the old jetty at Tokaanu or the dead tree by the Waihi marshes they will usually be found to have fed on carp, if shot at the Old Orchard at the mouth of the Waikato River, Taupo, Galaxias, Mototere Point, Gobiomorphus, and so on, so that the percentage of the birds that live on any definite food could be varied to a large extent by the site where the shooting takes place.” Acknowledgments. Thanks are due to Dr. J. Scaife Armstrong for his permission to publish his Taupo results, and to the Waimate Acclimatisation Society for responding to a request for shags for examination. First Locality Group. Locality: Lake Taupo, near Waikato River outlet. Dates: Various between May, 1929 and November, 1936. Collector: J. Scaife Armstrong. Recorder: J. Scaife Armstrong.

(a) Black Shags (P. carbo). All adult. Number of birds examined, 15, of which only 7 contained food. No. of birds in which found. Highest number in any one bird. Total from all birds. Food organism. Fishes— Salmo sp. 3 3 5 Galaxias sp. 1 1 1 Carassius sp. 1 1 1 Unidentified 2 — — Plants— Green Weed 1 — 1 (b) White-thrcated and Little Pied Shags (P. brevirostris). Number of birds examined, 40, of which 35 contained food. Fishes— Galaxias sp. 6 6 12 Gobiomorphus sp. 14 8 63 Carassius sp. 2 2 3 Unidentified 9 — — Crustacea— Paranephrops 6 4 12 Insects— Trichoptera, larvae 1 — — The series of P. carbo is not large enough to give a reliable indication either of diet range or food preference. It may be remarked that a high proportion (8 out of 15) had no food in the alimentary tract when shot, indicating that feeding is not a continuous process, although of course there is the possibility that such birds may have been taken early in the day before feeding had commenced. The record for P. brevirostris is based on a better series and it indicates a strong preference for Gobiomorphus followed by Galaxias and Koura (Paranephrops). There are no trout over the whole period of seven years in an area where they are abundant. The record also includes the first reference known to us of the finding of the larvae cases of Caddis Fly in the stomach contents. Second Locality Group. (Black Shag.) Locality: Lake Waikare, Waikato. Date: September, 1931. (Time, between noon and 3 p.m.) Collector: A. T. Pycroft. Recorder: R. A. Falla. Number of birds examined, 11 adults, all of which contained food. No. of birds in which found. Highest number in any one bird. Total from all birds. Food organisms. Fishes— Anguilla sp. 11 3 20 The birds were feeding young in a nesting colony, and the figures are of interest only as indicating a local preference where eels are available, and other food probably not so easily accessible. Method in Later Samples. The method of examination employed in the preparation of the following tables differs from that of Serventy (1938, p. 298) in so far that the whole of the fragmentary material representing the remains of past meals has been subjected to analysis. When fishes occur in the stomachs it is seldom that a complete specimen or even a considerable portion thereof is present, and the following deter-

minations are carried only so far as the actual material permits. A single bone or scale may be specifically identifiable, and where this is so the species is recorded, but on the other hand a much greater quantity of material may not justify more than a family determination. In some cases only the order to which a fish belongs may be determined, and when not even this is possible specimens are recorded merely as fishes. In essaying any interpretation of these restricted determinations it is necessary to give consideration to the locality from which the material was obtained. A fish of the order Isospondyli taken at Lake Ellesmerc might be a Galaxiid, a Salmonoid, a Clupeoid, or a smelt, but if taken at the Wilberforce River would necessarily belong to one of the two former groups. An Eleotrid from Lake Ellesmere would almost certainly belong to the genus Gobiomorphus and in all probability would be G. basalis, but one from upland waters would most likely be Philypnodon breviceps. It is also to be noted, when dealing with nestling birds, that parts of a single fish may have been fed by the parent to each of several nestlings, and the species thus given a false value in the tables. In determining insects and other animals everything identifiable has been counted. The figures, particularly of caddis larvae cases, must be regarded as approximate, as many fragments are not readily distinguishable as the partly digested remains of a complete case or as separated portions. Third Locality Group. (Black Shag.) Locality: Lake Ellesmere. (The first 6 birds taken at Birdlings Flat flying to cliff nests from Lakes Ellesmere and Forsyth; time, between 3 p.m. and 5 p.m.; remainder taken at mouth of Selwyn.) Collectors: E. F. Stead and D. Hope. Recorder: G. Stokell. Adults. Serial Number. Food organisms and other material. 1 Anguilla sp. 1 2 Rhombosolea sp. 1 Electrid fish 1 Isospondyli fish 1 Mollusca (Corneocyclas) 5 3 Piece of felt. Apparently remains of a gun wad 4 Anguilla sp. 1 Traces vegetable matter 5 Anguilla australis 1 6 Anguilla australis 3 Anguilla sp. 3 7 Gobiomorphus basalis 1 Eleotrid fishes 8 Pairs of otoliths similar to those of Eleotrids 39 Waterweed (Elodea canadensis) 8 Gobiomorphus basalis 6 Eleotrid fishes 4 9 Eleotrid fishes 4 Pairs of otoliths similar to those of Eleotrids 23 10 Pairs of otoliths similar to those of Eleotrids ½ Mollusca (Potamopyrgus) 4 Trichoptera, larvae 5 Seeds of (?) 7

11 Pairs of otoliths similar to those of Eleotrids 2 Hemipteia, adult 1 A Summary of these data is given below. Number of birds examined, 11 adults, 10 of which contained food. Food organisms. Fishes— No. of birds. in which found. Highest number in any one bird. Total from all birds. Anguilla 4 6 9 Rombosolea sp. 1 1 1 Eleotrids* Figures given in Summary take no account of large accumulation of otoliths. 4 10 24 Isospondylid Fish 1 1 1 Molluscs— Potamopyrgus sp. 1 4 4 Corneocyclas sp. 1 5 5 Insects— Trichoptera, larvae 1 5 5 Plants— Elodea canadensis 1 — — Seeds of (?) 1 7 7 Fourth Locality Group. (Black Shag.) Locality: Waitaki River. Date: November 25, 1938. Collector: Waimate Acclimatisation Society. Recorder: G. Stokell. Adults. Serial Number. Food organisms and other material. 1 Philypnodon breviceps 1 Salmo trutta 1 Archichauliodes dubitatus larvae 21 Nestlings. 2 Archichauliodes dubitatus larvae 1 Trichoptera, larvae 6 Plectoptera, larva 1 Pieces of Rock 4 3 Salmo trutta 1 Salmonoid fish 1 Trichoptera, larvae 824 Coleoptera, adult 1 Unidentified arthropod (fragments) Traces vegetable matter Piece of bark 1 Pieces of coal 3 Pebble 1 4 Salmo sp. 1 Trichoptera, larvae 263 Unidentified arthropod (fragments) Piece of wood 1 5 Isospondylid fish 1 Trichoptera larvae 438 Unidentified arthropod (fragments) Piece of rock 1 6 Trichoptera, larvae 602 Coleoptera, adults 6 Hemiptera, adults 3 Unidentified arthropod (fragments) Pebbles 7 Sand and grit

7 Salmo trutta 1 Salmonoid fishes 2 Trichoptera, larvae 31 Diptera, larvae 1 Coleoptera, adults 2 Archichauliodes dubilatus, larvae 3 Willow leaf 1 Gorse bloom 1 Seed heads of thistle 2 Other vegetable matter Sheep dropping 1 Piece of coal 1 Grit Summary. Food organism. Fishes— No. of birds in which found, Highest number in any one bird. Total from all birds. Salmo trutta or Salmonoid fishes 4 3 7 Philypnodon breviceps 1 1 1 Isospondylid fish 1 1 1 Insects— Archichauliodes dubi- tatus larvae 3 21 25 Trichoptera, larvae 6 824 1900 Plectoptera, larvae 1 1 1 Diptera, larvae 1 1 1 Coleoptera, adult 3 6 9 Hemiptera, adult 1 3 3 Miscellaneous material already listed above 6 — — This sample, although small, is of interest as representing the food of nesting birds. The quantity of insect larvae is striking, and the question arises as to its origin, whether gathered directly or as stomach contents already in fish captured. The latter suggestion, while tenable in respect of some birds, is strongly discredited by several circumstances noted in others. Caddis larvae are usually associated in the stomachs with the remains of fishes, and when the number of the former is consistent with the size of the latter secondary derivation must be considered possible, but the tables provide instances of stomachs containing a quantity of caddis cases that could not have been derived from the fishes present. Disregarding the data from nestlings, which may not give a true indication of the feeding habits of the parent birds, it may be noted that the single fish of about five inches in total length which occurred in No. 5, Group 5, could not have contained the 500 caddis with which its remains were associated. In some specimens the caddis cases were in the lower part of the stomach and in an advanced stage of digestion, but in others they were in the gullet and in so fresh a state that the animal could be removed from the case in perfect condition. It is scarcely conceivable that such larvae could have been in the shag's stomach long enough to have been digested out of a fish. In waters such as the Upper Selwyn and the Wilberforce, Black Shags may be noted wading in shallow ripples, and frequently submerging their heads as if grubbing under stones. These shallow rapids are the habitat of numerous caddis larvae and of varying

numbers of all other types of aquatic larvae enumerated in the tables. They also harbour fishes, principally Eleotrids and Galaxiids of small size. At the time Group 5 was being collected shags were observed working the shallow ripples of the Wilberforce as described above. Whatever the nature of the food that was being taken it does not appear to have been fishes as the birds collected that day contained no Eleotrids and only a single determinable Galaxiid, the latter being larger than usually occurs in the open river. In Group 4, which contained numerous caddis larvae, no Galaxiids and only a single Eleotrid were present. The circumstances, therefore, strongly suggest that where shags are working shallow, rapid water they are gathering caddis and other larvae directly from the stream bed. The size of the pebbles and pieces of rock taken from some shag stomachs further discredits the suggestion of secondary derivation. Stones are found occasionally in trout that have been feeding largely on caddis, the indication being that where several cases are attached to a small stone the whole mass is swallowed, but these stones are smaller than those contained by shags. Twenty-five per cent, of the latter are about ¾ inch in diameter, but it is very seldom that a stone of this size is found in trout, and it is to be noted that this applies to trout ranging from 9 inches to 16 inches in total length, whereas the fishes found in Groups 4 and 5 had a maximum of 6 inches. The ingestion of a ¾ inch stone by a 6 inch trout would appear to be a physical impossibility. Fifth Locality Group. (Black Shag.) Locality: Wilberforce River, Canterbury. Date: October 29, 1939. (Time, between noon and 3 p.m.) Collectors: G. Stokell and members of the Methven Gun Club. Recorder: G. Stokell. Adults. Serial Number. Food organisms and other material. 1 Salmonoid fishes 2 Trichoptera, larvae 38 2 Salmonoid fish 1 Trichoptera, larvae 137 3 Salmonoid fishes 3 Trichoptera, larvae 38 4 Salmonoid fishes 3 Flying Young. 5 Isospondylid fish 1 Trichoptera, larvae 500 Coleoptera, adult 1 Fragments of wood and roots Piece of clay 1 Piece of peat 1 Much sand and grit up to pieces ¼ in in length 6 Fish 1 Trichoptera, larvae 42 Hemiptera, adult 1 Mollusca (Potamopyrgus) 1 Traces vegetable matter Pieces of rock up to ¾in in length 16 A little sand and grit

7 Fish 1 Trichoptera, larvae 37 Mollusca (Potamopyrgus) 8 Mollusca (Isidora) 2 Waterweed (small quantity) Pieces of rock from 3–16in to ¾in in diameter 25 Nestlings. 8 Salmo trutta 1 Salmonoid fishes 2 Trichoptera, larvae 148 Archichauliodes dubitatus, larvae 2 Gordius worms, larvae 3 Fragments vegetable matter 9 Salmo gairdnerii 1 Trichoptera, larvae 157 Archichauliodes dubitatus, larvae 2 Pebble 1 10 Trichoptera, larvae 127 Archichauliodes dubitatus, larvae 2 Piece of wood 1 Piece of bark 1 Twigs, grass and other vegetable matter Pebble 1 Sheep droppings (globular) 6 11 Salmonoid fish 1 Trichoptera, larvae 349 Archichauliodes dubitatus, larvae 3 Coleoptera, adult 1 Diptera, larvae 1 Sheep dropping 1 Pebbles 8 Piece of wood 1 Piece of peat 1 Traces vegetable matter 12 Salmonoid fishes 3 Isospondylid fish 1 Trichoptera, larvae 169 Archichauliodes dubitatus, larvae 2 Coleoptera, adult 1 Plectoptera, larvae 1 Sheep dropping 1 Traces vegetable matter 13 Galaxias sp. 1 Salmonoid fish 1 Trichoptera, larvae 50 Stenoperla prasina, larva 1 Shot 1 14 Salmonoid fishes 2 Trichoptera, larvae 89 Archichauliodes dubitatus, larvae 2 Procordulia, smithii, adult 1 Roots 15 Salmonoid fish 1 Fish 1 Trichoptera, larvae 276 Archichauliodes dubitatus, larvae 4 Coleoptera (Odontria), adult 2 Diptera, larvae 2 Unidentified arthropod (fragments) Fragments of roots Pebbles 9 Sheep dropping 1

16 Isospondylid fish 1 Trichoptera, larvae 75 Diptera, larvae 1 Insect larva 1 Pebbles 2 Sheep dropping 1 Fragments of roots and moss 17 Salmonoid fishes 2 Trichoptera, larvae 276 Diptera, larvae 2 Plectoptera, larva 1 Coleoptera, adult 3 Archichauliodes dubitatus, larvae 3 Unidentified arthropod (fragments) Sheep droppings 3 Pebbles and splinters of stone 11 Pieces of wood 2 Piece of bark 1 Leaves of rata 2 Leaf of (?) 1 Moss and roots 18 Salmo trutta 1 Isospondylid fishes 2 Trichoptera, larvae 83 Diptera, larva 1 Coleoptera, adult 1 Sheep droppings 3 Pieces of wood 2 Pieces of bark 3 Pieces of rock. Largest ⅝in in length 2 Fragments of moss Summary. Number of birds examined: 4 adults, 3 fully fledged flying young, 11 downy nestlings (total, 18 birds, all containing food). Food organism. Fishes— No. of birds. in which found, Highest number in any one bird. Total from all birds. Salmonoid fishes 13 3 24 Galaxias sp. 1 1 1 Isospondylid fishes 4 2 5 Undetermined fishes 2 1 2 Insects— Trichoptera, larvae 17 500 2594 Archichauliodes dubi- tatus, larvae 8 4 20 Procordulia smithii, ad. 1 1 1 Stenoperla prasina, larva 1 1 1 Plectoptera, larvae 2 1 2 Diptera, larvae 4 2 4 Coleoptera (including Odontria) 6 3 9 Molluscs— Potamopyrgus sp. 2 8 9 Isidora sp. 1 2 2 Miscellaneous matter listed above 14 — — These data agree so closely with those of the previous group, and the circumstances of locality and class of bird are so similar that the combined results may be regarded as giving a safe indication of the food of nestling birds on upland waters.

Conclusions. The results listed above may be regarded as giving an estimation of the proportional composition of shag food (on a numerical basis) over a wide area and at various seasons. It yet remains to record the annual cycle from any one area. The numerical basis of presenting results, although necessary as a first step and valuable in itself, is not finally sufficient for applying the results to an evaluation of the Black Shag in the ecology of lakes and streams. Either the volume or the weight of the various food organisms must eventually be considered. In practice, owing to varying degrees of digestion, direct measurement of stomach contents is unsatisfactory, and the indirect method recommended by Serventy (1938, p. 297) is to be preferred. He writes: “I have attempted to avoid the difficulty by recording, during the course of my investigations, the lengths (or probable lengths in the case of partially-digested organisms) of the various food species as they were counted, and it had been my intention to calculate their original weights from data obtained from lengthweight ratios of fresh material.” This has not been done in the present paper, but could be applied to the data as set out once the necessary length weight ratios had been worked out. In view of the great mass of published material that has appeared in the daily press, in monthly journals and in annual reports purporting to be investigation, it seems desirable to point out some ways in which results of a carefully detailed analysis of stomach contents may modify conclusions more hastily reached. As an example of a moderate and well-intentioned report we may cite the following extract from the New Zealand Herald, 3/12/38, which records a local examination, of a sample from Locality 4, taken at the same time and from the same colony as the seven birds examined by us: “In every case the birds were fully gorged with fish, evidently trout. In some cases the fish, which were swallowed whole, were seen to be of takeable size. “Mr. Cox, in commenting on the results of the drive, said that it was a telling refutation of the charge made by a northern society that the Waimate Society was convicting the shag of feeding on trout on slender evidence. ‘When it is realised that the daily diet of a mature shag is about 2 ½lb of fish,’ he said, ‘and that at a low estimate there are about 800 shags between the dam and the river mouth, it is a wonder there are any fish in the river at all.’” The first discrepancy is in the finding “fully gorged with fish, evidently trout,” as compared with the more mixed diet indicated in the analysed sample. On the subject of the weight of food taken per day it may be remarked that although 2 ½lb is the lowest of such popular estimates that have come to our notice, its accuracy is open to question. It is difficult to estimate under open field conditions, but it is legitimate to point out that controlled experiment indicates that the average amount of food taken daily is the amount actually required, and that this average does not vary with conditions of

available food. Wetmore (1927), using caged cormorants in an enclosure, found that they were maintained in healthy breeding condition on an amount of fish varying from ¾lb to lib per day for six days a week. Although unable to experiment thus with shags we have been able to do so with penguins, and find ½lb to be the optimum daily requirement of a bird weighing 2¾lb. On the same rates a black shag's daily requirement would be about lib. Except in the case of some birds containing eels, the actual or estimated amounts pf food found in any one stomach did not exceed this in any of the samples examined in the present investigation. Varying degrees of digestion down to practically empty stomachs are found, indicating that feeding is restricted to a portion only of the daylight hours, as field observers have already recorded (Stead, 1932, p. 9). From the lists and analysis presented in this paper it should be possible to proceed to the third requirement—viz., an assessment of the place of shags in the whole ecology of lakes and streams. It may be pointed out that the motive of the present work is not vindication of the shag so much as a vindication of scientific method in approaching the question. References. Falla, R. A., 1937. The Distribution and Feeding Habits of Fresh Water Cormorants in New Zealand. Rep. Twenty-third Meeting Aust. and N. Z. Ass. Advancement of Science, p. 136. Serventy, D. L., 1938. The Feeding Habits of Cormorants in South Western Australia. The Emu, xxxviii, Pt. 3, pp. 293–316. Stead, E. F., 1932. The Life History of New Zealand Birds. Search Publishing Co., London, p. 9. Wetmore, A., 1927. “The Amount of Food Consumed by Cormorants.” The Condor, vol. xxix, pp. 273–274.