Studies in the Development of the Rainbow Trout (Salmo Irideus): I. The Heat Production and Nitrogenous Excretion

Animal development can be studied as a transfer of energy from non-living unorganized reserves to a living organized embryo. The egg is a small amount of protoplasm with a supply of supplementary material—the yolk—which varies considerably in amount in different kinds of eggs and from which there arises an embryo containing food reserves insignificant in amount. It is possible to draw up a balance sheet of the energy reserves in the system and the energy lost by the system between the beginning and the end or between any two arbitrary points in time throughout development. This is attempted in the present study of the development of the trout embryo from hatching until the end of the yolk-sac period when, under normal conditions, the alevin begins to eat.

Earlier studies of the energy metabolism of the embryo are reviewed in Needham (1931, §7). Heat production during a substantial proportion of the period of development was measured for the chick by Bohr & Hasselbalch (1900,1902) and in the toad by Gayda (1921). Bohr & Hasselbalch’s work is alone comparable with the present study in that the heat production observed was related to the metabolism. In the chick the main energy source is fat and deductions about its metabolism could be drawn from measurement of oxygen uptake and carbon dioxide output. Needham attempted an extended correlation between chemical and energy metabolism largely based upon this work, that of Murray (1925) and his own observations. Murray had concluded from his fuel values of the chick embryo that, in the early stages at least, the usual coefficients for relating chemical analysis and fuel value did not obtain. Needham, accepting this conclusion, was able to outline the framework of development for both chemistry and energetics, but to do this observations of several workers had to be extrapolated and adjusted. It has been the aim of the present study to accomplish sufficient experimental work on the same or similar batches of material by the same person in the hope of securing a more unified and consistent picture of the relationship between the chemistry and the energetics of development.

It will be shown that during the whole cycle of development in the trout the usual constants which relate (1) nitrogen content to protein content, (2) protein to fuel value and (3) fat content to fuel value can be applied within the limits of error of the several techniques used in the investigation.

Observations were begun in the breeding season of 1933 and continued for the four succeeding years, the last sequence of measurements relating to the period December 1936 to March 1937. In all, nine batches of eggs were investigated, and with each repetition, the range of observation was extended. Much reduplication has resulted from this method of repetition with amplification, and, for publication, only those data relating to the last batch of eggs, upon which the most complete series of observations were made, will be used. Certain observations made on earlier batches have been used for interpreting results obtained later from other materials—a procedure which may be criticized. Conditions of rearing were as far as possible identical for all batches. The complete tables of observations of work up to March 1936 are given in the Thesis for the Ph.D. degree deposited in Cambridge University Library.

The series of measurements made upon the developing egg-alevin system were:

• Upon the living material:

  • (a)

    the rate of heat production—sometimes daily, invariably every second day—from the first day after fertilization to the end of the yolk-sac period ;

  • (b)

    the rate of ammonia and urea excretion over roughly the same period as the heat production measurements.

• Upon fresh material. Determination of :

  • (c)

    wet weight of embryo and yolk ;

  • (d)

    dry weight of embryo and yolk;

  • (e)

    total carbohydrate content of the whole system.

• Upon dried samples from II (d). Determination of:

  • (f)

    the fuel values of embryo and yolk by combustion in a bomb calorimeter;

  • (f)

    the total nitrogen content of embryo and yolk samples ;

  • (h)

    the total material extractable by carbon tetrachloride.

The determinations I (a) and (b) and II (c), (d) and (e) were made over the same period on the same batch of eggs. The results of analyses in § III are used to interpret the 1936–7 observations. Since a large quantity of material is necessary for the combustion in § III (f) a large batch was reared specially for the purpose.

The trout egg was selected for this study for several reasons :

1. eggs can be obtained in relatively large numbers from one female;

2. eggs can be fertilized almost simultaneously;

3. eggs are easily reared and survive sorting into weight groups;

4. greater uniformity in development may be secured by selecting those which hatch within a limited period;

5. moribund eggs can be detected with ease;

6. the embryo does not feed for about eighty days when reared at 10 ° C. ;

7. the embryo can be separated by dissection from the yolk-sac and without contamination with yolk. (Since embryos are small, large numbers are required for a single determination; this minimizes the effect of sampling errors.)

The rainbow trout (Salmo irideus) eggs were reared in hatchery trays kept in a constant temperature room at 10 ± 0·5°C., supplied with a continuous flow of water maintained at a temperature of 10 ± 0· 2° C. for the 80-odd days of the experiment. The water supply came from a tank at constant level and its temperature was recorded continuously. The eggs were sent ‘green’; that is immediately after fertilization, from the Surrey Trout Farm Hatchery at Nailsworth, Glos. Mr Steven of the hatchery selected the eggs from one female and reserved samples of the same eggs unfertilized. From the fifth to the twelfth day any shock disturbs gastrulation overgrowth: only part of the yolk is surrounded. This results in the death of the embryo. Once overgrowth is complete, however, deaths are infrequent until hatching, at which time some die. After hatching, alevins are negatively phototactic and must be reared in complete darkness otherwise they kill themselves in an endeavour to swim through the hatchery trays.

As they form the connecting link between all analyses on dried samples and observations on the living egg and alevin, the determinations of wet and dry weights are fundamental to subsequent treatment. Unless the contrary is indicated all observations at whatever stage in development apply to the system embryo + yolk-sac = alevin, and not to the egg. At as early a stage as possible before hatching the alevin was removed by dissection from the egg membrane or ‘chorion’. This operation is facilitated if the eggs are hardened in 4 % formalin before dissection ; but comparisons showed that the water content of the embryo may be changed to a significant degree by such fixation. Accordingly all figures quoted refer to the fresh unfixed material.

Wet weights of groups of five alevins were determined in the following way. The material was dried superficially on a Buchner funnel in a current of air after rinsing rapidly with 96 % alcohol, and transferred to an open watch-glass. At a fixed period of time from the rinsing, the weight was taken. They were then washed with water on to a clean porcelain tile, and the yolks separated by dissection using cataract knives in the water drop. Under these conditions the yolk rounds off cleanly. No attempt was made to weigh the yolks directly ; their weight was obtained by difference. The dissected embryos after being dried superficially and weighed as before, were then dried in a porcelain dish to constant weight in an electric oven at 100· 5ଁ C. The yolks were left to dry on the tile, removed as fine flakes by scraping, transferred to small porcelain dishes, dried as before, cooled in a desiccator and weighed as rapidly as possible. Both dried embryo and yolk material are very hygroscopic.

The technique of heat production measurement was a modification of that of Hill (1911). Each microcalorimeter consisted of two selected vacuum flasks of approximately 200 ml. capacity with identical characteristic for heat conduction for a water content of too ml. One flask (the control) contained water alone while the other contained the experimental material with water sufficient to make a total volume equal to that of the contents of the control flask. The temperature difference between these flasks is measured by a thermocouple over some period of time, during which period production of heat in the experimental flask leads to a steady drift in the temperature difference between the flasks. The couples were calibrated against Beckmann thermometers to the nearest ⁠. Work on these lines was done for the three years 1934–5, 1935–6, 1936–7, with apparatus which was repeatedly modified; the description which follows refers to the final form (1936–7). The nature of the material necessitated that (i) measurements be conducted at 10° C., some 5·6° C. below normal room temperature, and (ii) a continuous supply of air for aeration be maintained. The flasks were fitted with glass collars so that the entire flask could be held below the surface of the bath. A stream of saturated air at the temperature of the bath was passed down glass tubes fitted with sintered glass filters to generate a stream of fine bubbles into each flask. Around these tubes three sets of soldered copper-constantin thermojunctions were constructed: two two-junction couples to record the temperature difference between each flask and the bath, and a four-junction couple to record the temperature difference between the flasks. Each tube thus carried twelve metal wires, and to reduce heat exchanges by conduction those copper wires which dipped into water were 42 s.w.g. copper wire. All couples were insulated with four coats of Parson’s Bakelite yacht varnish (no. 27) kindly supplied gratis by the firm, each coat being baked to hardness at temperatures rising from 60 to 120° C. over a period of 9 hr. The couples and intake tubes were held steady in the flasks by loose wads of cotton-wool in the glass collars. The rate of aeration was adjusted to 1–2 bubbles a second as observed by bubble counters. Experiments were always performed in duplicate. The e.m.f. of the couples was measured by a short-period moving coil galvanometer of low resistance with an auxiliary all-copper resistance box in series. All switches and coils were of copper. The galvanometer was calibrated against a Weston standard cell and a potential divider giving 10⁻⁵ of the input. To avoid errors from shifting of the galvanometer zero, and to gain greater sensitivity reversal throws were measured. The optical system was arranged to work at 4 m. instead of at 1 m., the standard working distance. Care was required in handling the apparatus, as stray heating effects in the circuit may cause temporary and transient fluctuations larger than the actual measurements recorded. Experiments were rarely run for more than 10 hr. as the rate of heat production tends to fall off after this. In order to estimate the extent of heat losses, and so to correct the heat-production curves, the conductivity of the experimental flask was determined both for heat losses and heat gains under working conditions at intervals throughout the course of the work. The conductivity factors were consistent to within 12–14% for each apparatus, and the means of all determinations were used in correcting the heat-production curves. The application of corrections for flask conduction followed Hill very closely. In no case did the total correction exceed more than 25 % of the observed values. The error in a single experiment was well within 10%.

The characteristic conduction (K) is determined by a number of factors:

1. conduction through the flask wall (identical in both flasks at 100 ml. content) ;

2. exchange of heat between the air aboye the bath and the flask contents by conduction along the leads (if the couples are symmetrical the differential effect is minimal);

3. the heating effect of the air current. This is difficult to control. The rate of bubbling is maintained as constant as possible, whilst heat losses from evaporation are reduced by aerating with saturated air. Since a pressure excess is required to produce air bubbles 3 cm. below the surface of the water in the flasks a small cooling effect is produced by aeration.

Each determination was made as follows. Fifty alevins were made up to a total volume of 100 ml. in a measuring cylinder and transferred to the experimental flask. Using the couple itself as recorder, ice-water was added from a graduated pipette until the temperature was a little below that of the bath and the control flask (a few thousandths ° C.), the corresponding volume of Water being removed to keep the heat capacity and volume of the system constant. Couples were adjusted, wads of cotton-wool fitted, the rate of bubbling adjusted, the lid put over the thermostat bath, and the whole left for 30 min. to settle down. Readings were taken at convenient intervals for 8–9 hr. The temperature-time curve obtained was corrected for heat losses. Knowing the heat capacity of the system, the sensitivity of the galvanometer and the characteristics of the couple the absolute heat production may be calculated.

The fuel values for dry embryo and dry yolk were determined by combustion in the standard Berthelot-Mahler bomb calorimeter. In this case the size of sample required for the analysis was large (more than 0 · 5 g.) and so a special batch of eggs was reared for these determinations. The dried yolk was of necessity burnt as a powder because on pressing it into a pellet fat is squeezed out. The apparatus was calibrated using benzoic acid. Experimental errors lay within 0 · 2%.

The method used for estimating ammonium ion and urea in dilute solution was developed from that given by Teorell (1932) for the estimation of ammonia in distilled water. Advantage was taken of the absorption of ammonia by permutit (Folin & Bell, 1917) in order to remove ammonia from solutions of ammonia and urea.

Each reaction is complete to within 0 · 5% for fixed concentrations of NaOH and Br₂.

Sodium hypobromite is added in excess of that required for the reaction and after a fixed period of time HBr is added and the bromine thus liberated is titrated with a N/2500 solution of naphthyl red (benzene-azo-α-naphthylamine). A second determination with NaBrO is done on a sample from which NH₃ has been removed by permutit.

The empirical values of k₁ and k₂ are determined for each lot of solutions and each sample of permutit and are controlled from time to time during experiments against pure solutions of ammonium sulphate and of urea and also of mixtures of the two. The method was used for quantities of 1–20 μg. of each in 10 ml. of water. Each determination took about 45 min.

A determination was made as follows : two samples of water, one before the eggs or alevins had been in many minutes, and a second after a recorded period of time. A sample of plain tap water was analysed also—it served as a blank control on the reagents used. The number of individuals was noted in each case. The interpretation of results of analyses was difficult when eggs died during a determination or when the alevins hatched or suffered damage to the yolk-sac. When deaths occurred numbers were ambiguous and the amino fraction was invariably increased by diffusion through the dead egg membrane of products not released by the living egg. Hatching liberated the syrupy vitelline liquid analysis of which therefore gave similarly enhanced amino-nitrogen values. Escape of yolk from the sac by damage leads to its denaturation and affects the amino-nitrogen values similarly. Two samples of 50 ml. of the water sample and of 50 ml. of tap water for blank are measured into 100 ml. graduated flasks. To each 20 ml. N/100 hydrochloric acid is added so that the resulting solution is very slightly acid (roughly 0 · 2 ml. 0 ·01 N HC1 in excess). To one sample 1 g. of permutit is added and shaken for 5 min. The flasks are made up to 100 ml. with ammonia-free glass-distilled water.

The procedure in titration is to withdraw 10 ml. samples into carefully cleaned 50 ml. Kavalier Erlenmeyer conical flasks (this glass is not tinted) covered with a glass cap. 2 ml. of the sodium hypobromite solution is added and the cover replaced, then shaken and left for 3min. (measured by stop-watch), 0 · 5 ml. 5 % hydrobromic acid is added and after 30 sec. naphthyl red solution is run in from a microburette until the pink colour of the dyestuff persists. The composition of the naphthyl red solution and the sodium hypobromite solution are given by Teorell (1932).

It is essential that the flasks be clean, and the several stages of the titration must be performed to fixed times as the volume of the naphthyl red required depends upon:

1. the duration of the titration;

2. the interval between the addition of hypobromite solution and acidification ;

3. the interval between acidification and titration.

Determinations of the total nitrogen in the dried samples of embryo and yolk were made by a microchemical modification of Kjeldahl’s method. Errors arising from limitations of technique lay within ± 0· 5 % for determinations using between 2–6 mg. of ammonium sulphate. Analytical fluctuations beyond this range on the trout material may be attributed to lack of homogeneity in the samples.

Details of the analytical procedure used in determining the total carbohydrate content of egg embryo and yolk-sac appear in the next paper of the series. For the argument of the present communication it is sufficient to state that the amount of carbohydrate present at any time and consumed during certain limited periods is so small (143 cal./100) in proportion to the total protein and fat consumption, some 10,000 cal., that in the general exposition of the balance sheet the carbohydrate content may be neglected.

Table 1 shows the wet weights in grams per 100 of egg, alevin and embryo and the yolk values calculated by difference. A study of these figures reveals no significant change in egg weight from fertilization to hatching. Until about the 40th day embryo growth proceeds slowly. The rate of weight increases slackens temporarily around the 50th day after which the rate of growth increases again until the 73rd day when a decline attributable to the using up of the yolk is observed. The alevin reaches its maximum weight before the embryo, a confirmation of Gray (1926). The wet weight of the whole increases as the intake of water by the system exceeds the loss in dry material resulting from catabolism. In the egg the restriction of water intake by the shell (‘chorion’) is a feature of development although Krogh and Ussing (1937) have observed that penetration of heavy water through the chorion is possible at all times and through the vitelline membrane after the eggs become ‘eyed’, some 15 days incubation at 10° C. The dry weights of 100 embryos and 100 yolks at various ages are also shown in Table I. The figure for the total weight of the system are obtained by adding the weight of the embryo and yolk respectively. This total decreases throughout the period studied, with a continuously accelerated slope corresponding to the increasing metabolic needs of the growing fish.

The measured rates of heat production expressed in calories produced by 100 individuals in an hour for the light batch of eggs of mean weight 90· 0 mg. (1936–7) are given in Table 1 and Fig. 1. Up to the 34th day eggs were used for the determination ; after this, alevins. The number used varied from too, when the anticipated heat production was low, to 25, so that irregularities due to population variability are not constant throughout the series. Figures for the heat production before the 15th day are only to be accepted with reserve. At this time several eggs died during each determination (up to a maximum of 80% between the 5th and 12th days), and it was difficult in the earliest stages to determine in the living egg how many were fertile. Not until the egg is ‘eyed ‘is it possible to ascertain fertility with certainty.

Within 24 hr. of fertilization heat production is negligible. Gastrulation overgrowth begins on the 4th–5th day and proceeds until the nth day under the conditions of incubation. The peak of heat production on the 4th day would then correspond to the end of segmentation. For short periods (5th–10th day) and before hatching at the 35th day the rate of heat production declines. It should be noted that until hatching was finished the agreement between duplicate measurements was less good than was to be expected from the physical characteristics of the calorimeters. As soon as the alevin hatched a marked increase in the rate of heat production was observed: this coincides with a marked increase in the rate of embryonic growth. The considerable fluctuations in heat production observed between the 65th and 83rd days have been confirmed during three seasons of observations.

The fuel values of egg, embryo, and yolk are given in Table 3. The samples used for these determinations were from a larger batch of eggs incubated for the purpose, on which heat production observations were not made. The mean weight of these eggs (66 mg.) was smaller than that of the 1936–7 batch (90 mg.), and the length of the yolk-sac period was shorter by some 3 days in 70. The time scale of development is influenced by size of egg and accordingly these fuel values were applied to the 1936–7 measurements by relating corresponding stages in the growth process instead of corresponding days of incubation. The acceleration in growth rate after hatching and the peak of embryo growth were selected for this purpose. The fuel values are used to calculate the total fuel value of the embryo, yolk and alevin in Table 4. Columns 2 and 4 show the dry weights of yolk and embryo respectively; column 5 the fuel value of 100 embryos, and column 3 the fuel values of the yolk at the same stage. The heat content of embryo and yolk are given in columns 6 and 7 respectively, the total heat content of 100 alevins is the sum of these, and is given in column 8.

The heat produced up to a given stage in development was calculated by numerical integration from the heat-production measurements. The values obtained in this way are shown in column 9 (Table 4).

The sum of columns 8 and 9 (Table 4) should be constant within the limits of the errors of determination of the quantities concerned if the first law is applicable to embryo systems. The disparity between the heat production observed and that to be anticipated from the fuel determinations is shown in Fig. 2 where the balance for the egg between fertilization and hatching and the beginning, middle, and end of the yolk-sac periods are represented schematically. The increase in integrated total heat evolved in development is shown by the increase in length of the upper section of the column ; the fuel value of the yolk is shown in the middle unit and that of the embryo in the lowest part. The discrepant parts of the balance are indicated by the hatched or shaded portions, these latter represent a fraction within the limits of error of the techniques employed and it is therefore reasonable to assume that the first law of thermodynamics is valid for the development of the trout embryo from egg to young fish.

The amount of ammonia and urea nitrogen excreted by the trout and alevin are given in Table 5; daily rate of nitrogen excretion is shown in Fig. 3. The hourly excretion of each 100 individuals rises, with short periods of arrested rates of increase at the 28th–29th, 37th–47th and 58th–6oth days. From the 75th day onwards there is a wide variability in rate of excretion and no general trend is discernible. The percentage of ammonia excreted by the egg is very high and there is no certain evidence of urea excretion by diffusion through the chorion before hatching. The slightest injury to an egg leads to the appearance of amino-nitrogen in the circumambient water. The few occasions on which urea excretion was negligible thus are taken to indicate the more probable state of affairs. The perivitelline fluid released on hatching contains large amounts of amino-nitrogen which seems unable to diffuse through the chorion. The perivitelline fluid of each egg produces on the average 37 μg. urea nitrogen. This may have accumulated as a product of the embryonic metabolism or as a result of digestion of the chorion by a hatching enzyme (Wintrebert, 1912) secreted by specialized cells in the ectoderm of the embryo.

The results of the total nitrogen analysis by micro-Kjeldahl methods are given in Table 6. The percentage of nitrogen in the yolk fraction changes considerably towards the end of the yolk-sac period, which changes may be attributed to selective absorption of proteins and to a proportionate increase of the remaining yolk fat. From this cause a decline in percentage nitrogen content occurs between the 45th and 70th days. The slight fluctuations in the embryo percentage nitrogen content while of statistical significance are not easy to interpret. The decline from the 75th day arises from the combustion of muscle protein by the starving alevin.

Table 7 was compiled using the results of dry-weight determinations for the development as an alevin from Table 1 and the percentage nitrogen figures of Table 6. Column 1 gives the age; 2 and 3 the figures to permit the calculation of nitrogen in the yolk-sac; and 4 and 5 similar figures for the embryo. The total nitrogen for the alevin is given in column 7, while column 8 gives the results of the total nitrogen excreted calculated by Simpson’s rule from the smoothed curve of total nitrogen excretion (Fig. 3). Column 9 is the sum of columns 7 and 8, and this total represents the total of nitrogen in the alevin system and of nitrogen lost by the system; this sum should be constant. The range of variation observed in column 9 affords a check upon the accuracy of the experimental techniques employed. Up to the 35th day a computation of the nitrogen balance sheet is difficult as only ammonia nitrogen is lost before hatching. This process entails the loss of urea nitrogen as the perivitelline fluid is released, while at the same time the material of the chorion has been dissolved by the hatching enzymes and the products of solution are absorbed by the alevin in considerable amount. An analysis of the chorions remaining after hatching was not made. These were difficult to collect and the viscid scum which they formed did not permit an estimate of their number. Careful analyses of the chorion were made some days before this, when it is a tough thick membrane ; after hatching it is thin and delicate. A sequence of weight readings from another egg batch when a number of chorions was allowed to dry to constant weight in air gave the following figures which are consistent between themselves being all from the same batch of eggs.

Hatching occurred at 34 to 36 days with this sample.

There is no doubt that a considerable solution of chorion occurs at hatching, and the evidence shows that such products of solution are absorbed to a considerable extent by the alevin system. From Tangl & Farkas’ (1904) figures 100 eggs contained 0·359 g. nitrogen while after hatching 100 alevins contained 0·357 g. nitrogen. Losses from the chorion were thus negligible. The balance between fertilized egg and hatched alevin are :

(The chorion figure in the left-hand column is calculated from Hayes’s (1930) figure of percentage nitrogen using his dry weight as a basis of comparison.) Throughout development the nitrogen of the egg may be considered accounted for, if due allowance be made for errors of experiment.

The substantial agreement between analyses of dry material and the observations on the living material, shown in the calories and nitrogen balances, invites some attempt to correlate these two sets of observations with the observed weight losses of the system. The procedure and argument are summarized in Table 8 and they are as follows:

The figure for cumulative nitrogen excretion in column 2 (up to the age given in column 1) are converted into estimated weight loss assuming the nitrogen excreted to be derived from protein catabolism and that the factor 6·25 ×N obtains for embryonic systems. The loss during hatching and the estimated nitrogen content of the chorion is omitted from this series of figures. These weight losses are reproduced in column 7 of the table. They are converted into estimated heat losses (column 3), by multiplying with the factor 4200 which is an average estimate of the fuel value of protein. The integrated cumulative heat losses of the system are given in column 4. Some part of each heat-loss figure may be ascribed to protein com bustion, the carbohydrate contribution is small and may be neglected and the difference column 4 minus column 3 therefore may be taken to represent the heat energy lost by the system through the combustion of fats. Column 5 gives such estimated figures. These may be converted into weight losses by dividing each figure by 9400 which is a typical fuel value for fat. The fat weight losses are given in column 6, and the protein weight losses in column 7. The total loss from fat and protein combustion is seen in column 8, being the sum of 6 + 7.

On the right-hand side of the table the dry weights of the alevins are given in column 12. The dry weight of the eggs is 3·121 g./100. Column 11 supplies the differences between this value and the dry weights of column 12 and each figure accordingly represents the observed loss in weight of the system. This of course includes the chorion and weight of the nitrogen substance released on hatching. Subtracting from this observed weight loss (column 11) the estimated metabolic losses of column 8, the ‘residues’ given in column 9 result. This residue is seen to fluctuate around a mean value of 0·115 g./100. This is a reasonable figure for the solids lost at hatching, which include the chorion residue and some of the products of solution of the chorion.

If all losses were correctly traced to metabolism and to chorion loss on hatching all these residues in column 9 should be constant, their fluctuation from the mean 0·115 is given as a difference in column 10. These differences are inconsiderable in amount compared with the possible errors in the dry-weight determinations, it is therefore reasonable to assume all weight losses in the system accounted for as loss in chorion at hatching and as fat and protein catabolized in the maintenance of the growing system. Fig. 5 expresses the results of this correlation, the lower curve gives the estimated cumulative loss in weight from protein catabolism, the upper the sum of the protein and fat losses. The crosses are the dry-weight losses less 0·115 hatching loss. There is clearly no significant difference between the curve deduced from the metabolism and the figures for dry-weight loss.

This paper presents the results of an attempt to observe, as completely and directly as possible, the energy changes occurring during embryonic development in the trout. It has been shown that the dry-weight losses of the system are in accordance with those calculated from the results of simultaneous measurement of heat production and of protein breakdown. A study of the carbohydrate metabolism to be described in a later paper has shown that for present purposes the energy changes associated with this fraction of the total metabolism may be neglected. It was found that the standard caloric equivalents for normal adult metabolism were also valid for the developing trout in respect to both protein and fat. There is therefore no essential difference between the thermal efficiency of embryonic and adult metabolism, within the limits of experimental error of the procedures employed. It is also clear that the first law of thermodynamics applies to embryonic development.

In the course of this analysis separate balance sheets for the heat energy of the system and its progressive dissipation, and for the total nitrogen of the system have been compiled. These show that all the nitrogen in the egg after fertilization is either excreted or converted into embryo. Similarly the total fuel energy of the egg is either dissipated or stored in the embryo. It seems relevant at this juncture to examine previous work within this field.

The most important work is that of Bohr & Hasselbalch (1902) on the metabolism of the hen’s egg. They made simultaneous measurements of the heat production and carbon dioxide output of single eggs, but were not able to make observations on a single egg throughout the entire period of incubation. From their results for different eggs of different ages they compile a graph which shows that the heat production observed corresponds to the heat production expected on theoretical grounds from the carbon dioxide production within an error of 4%. They assumed that all the carbon dioxide was derived from the metabolism of fats and that the normal calorie equivalent for fats applied. Though the early days of incubation were not studied in detail this work is the most considerable study yet published.

Gayda (1921) made extended measurements of the heat production of embryo and larval toads (Bufo vulgaris). He did not measure respiration nor did he perform chemical analyses. The heat production observed was correlated with empirical estimates of the surface area. For much of the period studied the larvae were feeding and living a normal active life. It was thus impossible to draw up a balance sheet relating energy intake to growth and to heat production. Gayda’s results were subjected to mathematical treatment by Wetzel (1937), for comment on this work Needham (1942, § 3·21) may be consulted.

In attempting to compile a balance sheet for the energy changes during development from his own and Murray’s studies of the chemical embryology of the chick, Needham (1931) experienced great difficulty in adjusting the results of several independent workers. His conclusion that: ‘it may, therefore, be assumed as probable that the true calorific constants for the substances present in early embryonic life (in the case of the chick) should be regarded as lower than those for the corresponding adult substances’ certainly does not apply to the development of the trout.

Until recently it was supposed that skeletal and muscular systems were permanent machinery which, though requiring periodic repair, did not need continual substantial replacement, and that a part of the energy requirement of growth was devoted to the establishment of organized systems of this kind. From the work on the rate of exchange of chemical molecules (using radioactive isotopes) (summarized in Rittenberg, 1941) in the organized tissue systems of the adult mammal, it now seems certain that a fairly rapid turnover of the constituent amino-acids of a muscle protein, for instance, occurs. Maintenance thus becomes of far greater importance than was envisaged by earlier workers on the energy relations of development.

Since orderliness can only be expressed in physical terms by invoking the concept of entropy (see Schrodinger, 1944), there has been a tendency in discussing the development of structural pattern in living things, to suppose that they evade in some way the limitations of physical laws. A developing organism, increasing in structural complexity, appears at first sight, to be an impertinent exception to that great generalization ‘Die Entropie der Welt strebt einem Maximum zu’. What is forgotten is that the entropy of the system as a whole is increasing as a result of the normal maintenance of living tissue and the exothermic breakdown of food materials to provide in sufficient amount the specific substrates necessary for development and growth.

A study of the energy balance between two arbitrary states, let us say the beginning and the end of embryonic development, can furnish no measure of entropy changes in some restricted part of the system. A local decrease in entropy will be concealed by the nett entropy increase of the system as a whole. Moreover, it seems extremely probable that local decreases in entropy (associated with the formation of fibrillar structures of various kinds, for example Picken (1940)) are intimately related to local syntheses. The inadequacy of entropy considerations alone and the importance of Gibb’s free energy has been pointed out by Butler (1946) who concludes : ‘if an organism can synthesize peptide bonds, it appears that it will have no great difficulty in putting together protein molecules of any degree of complication. The free energy must come from the metabolic processes going on in the organism…. There is thus no outstanding difficulty in accounting for the synthesis of living structures with a fairly modest expenditure of food.’

Until more is known of the energetics of chain molecule synthesis and fibrillar aggregation in biological systems we cannot expect to proceed further in the detailed study of the energetics of embryonic growth and development.

1. The wet and dry weights of a selected population of rainbow trout eggs were determined at intervals from the 1st day after fertilization until the 84th day of incubation at 10° C. The alevin attains its maximum weight at the 65th day, the embryo some 10 days later.

2. The heat production of egg and alevin rises until the 65th day and subsequently fluctuates violently. Fuel values of yolk and embryo samples were measured.

3. The heat production is commensurate with the loss in the total fuel value of the system.

4. Nitrogen excretion of eggs and alevin rises steadily throughout development. It seems probable that only the ammonia fraction is freely diffusible through the chorion.

5. Determinations of the total nitrogen content of embryo and yolk samples demonstrate that the nitrogen excretion observed is sufficient to balance the loss in total nitrogen of the system.

6. The fuel values and excretion measurements are shown to be in accord with the observed loss of dry materials by the egg-alevin system. The significance of this is discussed.

For the first two years of this investigation I received help from the Department of Scientific and Industrial Research. The Royal Society made two grants of money for apparatus. I was later a Research Fellow of St Catharine’s College. For all such aid I am extremely grateful.

The stimulation, advice and interest of Prof. James Gray and Dr Laurence Picken have been indispensable. Without the support of the assistants in the department of zoology this work would have been beyond the powers of one person. I wish to thank them for such aid, especially D. R. Ashby and R. S. Undrill.