Summary1. In analysing the ecological conditions of an animal population we have above all to focus our attention upon the most sensitive stages within the life cycle of the animal, that is, the period of breeding and larval development.2. Most animal populations on the sea bottom maintain the qualitatively composition of the species composing them, over long periods of time, though the individual species use quite different modes of reproduction and development. This shows that species producing a large number of eggs have a larger wastage of eggs and larvae than those with only a few eggs. The wastage of eggs in the sea is much larger than on the land and in fresh water.3. In the invertebrate populations on the level sea bottom, large fluctuations in numbers from year to year indicate species with a long pelagic larval life, while a more or less constant occurrence indicates species with a very short pelagic life or a non-pelagic development.4. In most marine invertebrates which shed their eggs and sperm freely in the water, either (a) the males are the first to spawn, thus stimulating the females to shed their eggs, or (b) an ‘epidemic spawning’ of a whole population takes place within a few hours. Both methods greatly favour the possibility of fertilization of the eggs spawned and show that the heavy wastage of eggs and larvae takes place after fertilization, during the free swimming pelagic life.5. Embryos with a non-pelagic development may originate (a) from large yolky eggs, in which case all the hatching young of the same species will be at the same stage of development, or (b) from small eggs which during their development feed on nurse eggs, when the individual embryos of the same species may vary enormously in size at the stage of hatching.6. Three types of pelagic larvae are known: (a) Lecithotrophic larvae, originating from large yolky eggs spawned in small numbers by the individual mother animals; they are independent of the plankton as a source of food although growing during pelagic life, are absent from high arctic seas but constitute about 1o% of the species with pelagic larvae in all other seas, (b) The planktotrophic larvae with a long pelagic life, originating from small eggs spawned in huge numbers by the individual mother animal; they feed from, and grow in, the plankton, constituting less than 5% of high arctic bottom invertebrates, 55–65% of the species in boreal seas, and 8o-85 % of the tropical species, (c) The planktotrophic larvae with a short pelagic life having the same size and organization at the moment of hatching and at the moment of settling; these constitute about 5% of the species in all Recent seas.7. To find out the factors which cause the enormous waste of eggs and larvae, we thus have to study those forms (constituting 7o% of all species of bottom invertebrates in Recent seas) which have a long planktotrophic pelagic life, as only species reproducing in this way have really large numbers of eggs.8. The food requirements of the planktotrophic pelagic larvae are much greater than those of the adult animals at the bottom. The adaptability of the larvae to poor food conditions seems, nevertheless, to be greater than hitherto believed. The significance of starvation seems mainly to be an indirect one: poor food conditions cause slow growth, prolong larval life, and give the enemies a longer interval of time to attack and eat the larvae.9. At the temperatures to which they are normally exposed, northern as well as tropical larvae seem on an average to spend a similar time (about 3 weeks) in the plankton. The length of the pelagic life of the individual species may, however, vary significantly in nature. In the Sound (Denmark) the larvae are never exposed to temperatures outside the range which they are able to endure. The wastage caused by temperature, like that due to starvation, seems mainly to be an indirect one: low temperatures postpone growth and metamorphosis, and give the enemies a longer time to feed on the larvae.1o. When a larva feeding on a pure algal diet metamorphoses into a carnivorous bottom stage, a ‘physiological revolution’ occurs and a huge waste of larvae might be expected. Experiments have, however, shown that this is not the case.11. Young pelagic larvae are photopositive and crowd near the surface; larvae about to metamorphose are photonegative. Larval polychaetes, echinoderms, and presumably also prosobranchs, may prolong their pelagic life for days or weeks until they find a suitable substratum. Forced towards the bottom by their photonegativity and transported by currents over wide bottom areas, testing the substratum at intervals, their chance of finding a suitable place for settling is much better than hitherto believed.12. Continuous currents from the continental shelf towards the open ocean may transport larvae from the coast to the deep sea where they will perish. Such conditions may (for instance in the Gulf of Guinea) deeply influence the composition of the fauna, while in other areas (European western coast, southern California) they seem to be only of small significance.13. The toll levied by enemies appears to be the most essential source of waste among the larvae. A list of such enemies, comprising other pelagic larvae, holoplank-tonic animals and bottom animals, is given on p. 2o. A medium-sized Mytilus edulis, filtering 1–4 1. of water per hour, may retain and kill about 100,000 pelagic lamellibranch larvae in 24 hr. during the maximum breeding season in a Danish fjord.14. Species reproducing in a vegetative way, by fission, laceration, budding, etc., might be expected to have good chances of competition in such areas where conditions for sexual reproduction are unfavourable. Nevertheless, they only supply a rather small percentage of the animal populations of all Recent seas, probably because their intensity of reproduction is low and because they are unable to spread to new areas. Most forms reproducing in a vegetative way have sexual reproduction as well.15. Pelagic development is nearly or totally suspended in the deep sea, and is restricted to the shelf faunas. In the arctic and antarctic seas pelagic development is nearly or totally suppressed, even in the shelf faunas, but starting from here the percentage of forms with pelagic larvae gradually increases as we pass into warmer water, reaching its summit on the tropic shelves.16. In order to survive in high arctic areas a planktotrophic, pelagic larva has to complete its development from hatching to metamorphosis within I–I ½ months (i.e. the period during which phytoplankton production takes place) at a temperature below 2–4o C. Most larvae, that is in 95% of the species, are unable to do so and have a non-pelagic development, but if a pelagic larva is able to develop under these severe conditions the planktotrophic pelagic life seems to afford good opportunities even in the Arctic. Thus the 5 % of arctic invertebrates reproducing in this way comprise several of the species which quantitatively are most common within the area.17. The antarctic shore fauna has poor conditions similar to those of the Arctic. The longest continuous periods of phytoplankton production are 2 and 3 weeks respectively, and pelagic larvae have, in order to survive, to complete their development within this short space of time at a temperature between 1 and 4o C. Accordingly, non-pelagic development is the rule, but most arctic species are able to support their non-pelagic development by means of much smaller eggs than the antarctic species, where brood protection and viviparity is dominant. The antarctic fauna has apparently had a longer time to develop its tendency to abandon a pelagic life. The greater the size of the individual born, the smaller its relative food requirements and the better its chance of competing under poor food conditions.18. The relatively few data on reproduction in deep sea invertebrates point to a non-pelagic development. The larvae of such forms, in order to develop through a planktotrophic pelagic stage, would have to rise by the aid of their own locomotory organs through a water column 2000–4000 m. high or more (often with counteracting currents) to the food producing surface layer, and to cover the same distance when descending to metamorphose and settle.19. The ecological features common to the deep sea, the arctic and the antarctic seas, which enable the same animals to live and to reproduce there, contribute to explain the ‘equatorial submergence’ of many arctic and antarctic coastal forms.20. In the tropical coastal zones where the percentage of species with pelagic larvae reaches its maximum, the production of food for the larvae takes place much more continuously than in temperate and arctic seas, because light conditions enable the phytoplankton to assimilate all the year round. The tropical species of marine invertebrates breed (in contrast to temperate and arctic species) within such different seasons that their larval stock, taken as a whole, is more or less equally distributed in the plankton all the year round. This makes the competition in the plankton less keen.21. The fact that a mode of reproduction and development, well fit for an arctic area, is unfit in a temperate or tropical area of the sea is probably one of the main reasons for the restricted distribution of species.22. In most groups of marine invertebrates the individual species have only one mode of reproduction and development, which accordingly restricts their area of distribution. In the polychaetes, however, the individual species often show an astonishing lability in their mode of reproduction and development which enables them to compete in wide areas of the sea. Thus, out of the Western European species of polychaetes, 28-4% have been found also in the Indian Ocean, and 18%, at least, along the Californian coast, while the corresponding number of Western European echinoderms, prosobranchs and lamellibranchs found also in the Indian Ocean and California amounts to less than 2%.23. The pelagic or non-pelagic development of marine prosobranchs has proved to be a very fine ‘barometer’ for ecological conditions. Recent observations, still not elaborated, seem to indicate that the shape of the top whorls, the apex, of the adult shells of prosobranchs may show whether the species in question has a pelagic or a non-pelagic development. This discovery may also give us valuable information about the larval development in fossil species, and help us to form an idea about ecological conditions in sea areas from earlier geological periods.