The Sensitivity of Barnacles and Their Larvae to Copper and Mercury

A number of workers (e.g. Powers (1917), Löhner (1924), Carpenter (1927, 1930), Ludwig (1927), Hykes (1931), Jones (1935, 1937, 1938, 1939, 1941, 1947), Miller (1946)) have investigated the effects of heavy metal poisons on a variety of animals, including Protozoa, Ctenophores, Platyhelminthes, Crustacea, Molluscs, Polyzoa and Teleosts. Cole (1932) published an account of the effect of a number of substances, some poisonous, on the rate of cirral beat of barnacles and, more recently, Müller (1940) and Clarke (1947) have added further to our knowledge of the toxic effects of copper and mercury on larval and adult barnacles.

Clarke’s investigation (1947) was carried out in connexion with an extensive investigation of the fouling problem, and the experiments described below also formed part of a fundamental study of the same problem. In the present instance experimental work has chiefly been carried out using the larvae and adults of Balanus balanoides, though some observations have been made on the corresponding stages of B. crenatus. As sufficient numbers of the larvae of both these species are only present in the plankton for short periods each year (Pyefinch, 1948b), experimental work involving these stages has been virtually limited to the month of April and this study has, therefore, been spread over some four years in all. It will be clear that much more information is needed before even the outline of this investigation can be regarded as complete, but this preliminary study is presented as a further contribution to what is already known about the effects of poisons on the acorn barnacle, undoubtedly the most important type of animal fouling organism.

‘Analar’ cupric sulphate (CuSO₄5H₂O) and mercuric chloride (HgCl₂) were used as sources of copper and mercury respectively. Stock solutions of these salts were made up in distilled water and the appropriate volume of the stock solution added to freshly filtered sea water for each experiment. The use of salts such as cupric sulphate severely limited the range of concentrations which could be used in natural sea water as, if the concentration of copper rises above 6–7 mg./l., precipitation of basic copper salts occurs.

Some experiments using this poison were therefore carried out in an artificial sea water of the following composition :

Distilled water to 1000 g. The solution was well aerated before use.

It is realized that the copper concentrations in natural sea water just quoted far exceed the true solubility of this element in this medium. The solution of copper in sea water is a complex problem, which is more fully discussed by Barnes & Stanbury (1948), and all that can be added here is that, up to concentrations of 6–7 mg. Cu/1., an increase in copper concentration in natural sea water produces an increased toxic effect. The means whereby this is produced are at present obscure.

No complications of this character occurred with mercuric chloride, and this reagent could be used in natural sea water over the whole range of concentrations employed in these experiments.

The concentration of poison present was determined either during or immediately after each experiment, using colorimetric methods. For the estimation of copper, the sodium diethyl-dithio-carbamate method was used (for details see Harris (1947) ; both the direct and the extraction techniques described in this paper were employed) and the amount of mercury present was estimated using the diphenyl-thiocarbazone (dithizone) technique (see Barnes, 1946; Harris, 1947). The transmittancies of the solutions were determined, by means of a Spekker Photoelectric Absorptiometer, fitted with Ilford spectrum violet filters (no. 601) for the copper determinations and with Ilford spectrum blue-green filters (no. 603) for the mercury determinations. In general, agreement between calculated and determined concentration was good.

For each experiment, the range of concentrations used was such as to cover the whole range between no apparent lethal effect and 100% lethal effect at the end of the exposure period which, with the exception of cases specifically noted, lasted 6 hr. The percentage killed at each concentration was plotted, and the median lethal concentration determined from this graph.

Further details of experimental technique are given in the later sections of this paper.

A small number of experiments was carried out on the effects of copper and of mercury on the later stage nauplii of B. balanoides and B. crenatus. These experiments were made merely to obtain an estimate of the sensitivity of these larval stages, for comparison with that of the cyprid and the adult barnacle ; an extensive survey of the effects of these poisons was not considered likely to be profitable.

In these experiments an aliquot portion of a suspension of the nauplii of the two species was added to the poison solution in small crystallizing dishes, so that the final volume of the solution was 50 ml. At the beginning of the exposure period in each concentration used the nauplii swam freely, and quickly took up a position either immediately below the meniscus on the side of the dish towards the light or at the bottom angle of the dish on the side away from the light. (These are the normal reactions of these larvae. Ishida (1936) has described similar responses for the nauplii of B. amphitrite albicostatus and Tetraclita squamosa and has termed them ‘skotophobic’ and ‘photophobic’ respectively; in the present instance it would seem that both responses could be due to a positive reaction to light, since a marked ‘lens ‘effect is present at the bottom angle of a cylindrical glass vessel.) Shortly before the end of the 6 hr. exposure period each dish was rotated through 90°; those larvae still capable of active movement took up their orientation afresh to the incident light; those moribund or dead remained in their old positions. At the end of the exposure period it was thus possible to remove the active and the inactive or moribund larvae separately. They were fixed in dilute formalin immediately after removal, and the numbers counted and the stages analysed (following the methods described elsewhere (Pyefinch, 1948a)).

The result of one experiment is shown in Table 1.

In each stage, mercury is more poisonous than copper ; it is of interest to note that, although the sensitivity of the nauplii of both species to copper does not change considerably as one larval stage succeeds another, there is a more marked change in the sensitivity to mercury. The ratio between the median lethal concentrations of the two poisons, therefore, decreases through the series.

These results cannot be compared directly with those of Clarke (1947) for the nauplii of Balanus eburneus, as, since his tests were made on the day the nauplii were released from the brood pouch, it is probable that only first and second stage nauplii were present, and also his experiments were continued for periods of 22,29 and 48 hr. However, from the data he gives (Clarke, 1947, table II), it is possible to estimate the median lethal concentration for the 22 hr. exposure period; this value is roughly 0·15 mg. Cu/1., which suggests that the sensitivity of the nauplii of B. eburneus is of the same order as that of the nauplii of B. balanoides and B. crenatus.

The experiments with free-swimming cyprids, which formed the major part of this investigation, were carried out in small glass bottles, each of approximately 130 ml. capacity. Duplicate determinations were usually made at each concentration of poison. Ten cyprids were placed in each bottle and, after the addition of the larvae, the bottle was filled to the brim and capped with a glass plate. The latter procedure was adopted because it was anticipated that the cyprids would become trapped in the surface film ; later it was found that this possibility was only likely to occur with the cyprid of B: crenatus, and as most of the experimental work with this stage was carried out with B. balanoides this precaution was not really necessary. The rate of oxygen consumption of the cyprid, however, is very low (of the order of 0·0007 ml. O₂/day)^* so that oxygenation of the test solutions was probably completely satisfactory, although the surface was sealed from contact with the air.

The cyprid, both of B. balanoides and of B. crenatus, is by no means continuously active, and a number of different methods were tried as indicators of activity at the end of the exposure period. That finally adopted was as follows. At the end of the experiment, the cyprids were removed from the bottle with a pipette and placed in a Petri dish. As much as possible of the poison solution carried over in this process was removed and the dish filled with fresh filtered sea water. The dish was then placed on a sheet of white paper marked with ten dots; each cyprid was brought into position over one of the dots and those which remained in this position over a period of at least 12 hr. after the conclusion of the experiment were counted as dead. Usually several inspections were made during this 12 hr. period.

The results of a number of determinations (made in 1947) show that the average median lethal concentration for the cyprid of B. balanoides is 5·9 mg. Cu/1., whereas the corresponding value for mercury (based on a smaller number of determinations) is 3·0 mg./l., giving a copper/mercury ratio of almost 2/1 (weight/weight). The free-swimming cyprid of B. crenatus differs markedly in its sensitivity to these two poisons ; only a small and variable fraction of the cyprids is killed in the concentrations of copper possible in natural sea water, so it is reasonable to assume that the median lethal concentration for this poison is greater than 7·0 mg./l., but it is much more sensitive to mercury than the cyprid of B. balanoides, as the median lethal concentration is appreciably less than 1·0 mg./l. For this cyprid, the copper/mercury ratio might well be of the order of 10/1.

In carrying out estimations of the sensitivity of cyprids of B. balanoides, it is essential to use comparable material. For these experiments, large numbers of the cyprids of this species were collected from the plankton at the time when they are present in considerable numbers, and stored in tanks in the laboratory. Settlement of the cyprids did not take place immediately, but usually occurred 4 or 5 days after catching. During this period the median lethal concentration, both for copper and for mercury, decreases; examples of the amount of this decrease for copper are given in Table 2.

Comparable decreases occur in the median lethal concentration of mercury so that, by selection of experimental results, it is possible to produce wide variations in the relative toxic effect of copper and mercury. Table 3 gives examples of some ratios that can be obtained.

The experiments which formed the basis of Table 3 were carried out in 1945, and it will be noted that, where comparable material (i.e. material which had been stored for the same length of time) was used, the copper/mercury ratio approximated to the value indicated by more extensive work 2 years later.

It seemed possible that these changes in sensitivity to these poisons (and particularly the change in sensitivity to copper, since there is a good deal of evidence which suggests that this element is closely bound up with intracellular processes and especially with intracellular respiration) might be an indication of physiological changes taking place during the life period of the cyprid. This suggestion received some support, on the basis of experiments made during 1945, from the variation in sensitivity to copper from day to day. These results are shown in Fig. 1. The median lethal concentration varies from day to day, and shows some correlation with the variation in numbers of larvae taken in the plankton each day and also, though this is not shown in the figure, with the rapidity with which these larvae settled under laboratory conditions. For example, the larval stocks from which samples were drawn for the sensitivity determinations made on 14 and 15 April began to settle in a time much shorter than the 4 or 5 days suggested as an average interval earlier in this account; the next day the sensitivity decreased, the numbers of larvae caught increased and the interval between catching and settlement also increased. One interpretation of this sequence of events is that the tests on 14 and 15 April were made on larvae which were approaching the end of their life period, whereas the test on the following day was made on cyprids which had recently moulted from the sixth stage nauplius.

This hypothesis was further tested by more extensive experiments carried out through April 1947; their results are shown in Fig. 2. This shows that the correlation between a decrease in sensitivity and an increase in numbers of cyprids taken in plankton hauls suggested by the determinations made in 1945 applies only to the closing period, and possibly also to a short period at the beginning of the 1947 experiments. A full correlation throughout the period of cyprid abundance in 1947 was not to be expected, as it is evident from Fig. 2 that the numbers of larvae caught day by day fluctuated considerably for much of the period over which sensitivity determinations were made. This suggests that the cyprid populations were, for some time, made up of a series of small batches of newly formed individuals. Unfortunately, there is no possibility, at present, of testing such a hypothesis by direct observation. Visscher (1928) found that the oil globules present in the tissues of the anterior end of the cyprids of B. amphitrite and B. improvisus disappeared as the time for settlement approached, but this change does not occur in the cyprid of B. balanoides. However, the relationship between sensitivity to copper, numbers of cyprids present in the plankton and the ease with which settlement occurred under laboratory conditions found in 1945, and confirmed over the later part of the readings 2 years later, suggests that variations in sensitivity to copper may be a valid indication of changes in physiological state.

It should be added that the relationship observed in 1945 is perhaps unusually clear, as daily sensitivity determinations were only made over the latter part of the period of cyprid abundance. Had these determinations been begun earlier, it is possible that the indications would have been less clear, because the cyprid stock earlier in the period of abundance of these larvae would have been composed, as it was in 1947, of a mixture of cyprids of all ages, some newly hatched and some ready to settle.

B. crenatus cyprids settle much more readily under laboratory conditions than the cyprids of B. balanoides and, so far as can be judged on the basis of a few experiments, there are no changes in the sensitivity of the cyprids of this species on storage comparable with those found for B. balanoides. This difference may merely be due to the fact that the conditions of laboratory storage are more natural for the cyprids of B. crenatus than for those of B. balanoides, since physiological changes in the latter must, it may be presumed, be markedly affected by the alternate immersion and exposure these cyprids encounter between tide marks.

Some estimations of the sensitivity of B. balanoides cyprids have been made in artificial sea water. The use of this medium enables a wider range of copper concentrations to be used and the results of the exposure of the cyprids of B. balanoides to such a range of concentrations in this medium are shown in Fig. 3. Lethal action increases with concentration up to 0·5 mg. Cu/1., but, from concentrations of this order up to 100 mg. Cu/1., lethal action decreases as the concentration increases, so that at the latter concentration, 90 % of the cyprids used recover when returned to sea water after a 6 hr. exposure period. If the concentration is increased further, the lethal effect again increases.

These results may be due to the use of artificial sea water, and thus cannot be compared with exposures using natural sea water. But the use of artificial sea water does not produce any striking differences in the action of mercury, and other (unpublished) work suggests that results obtained in artificial sea water are comparable with those in the natural medium. If these results are therefore truly due to the poison, and not to the medium, they suggest an interesting difference in the mode of action of copper and of mercury. It is not possible to elucidate the nature of this difference on the basis of the present series of experiments, but one possible explanation of the results obtained from exposures to copper in artificial sea water is that this element not only has a direct poisoning action which may be comparable with that of mercury, but also has an effect on some metabolic process which affects the rate of uptake of the poison. It should be added that, when tests are carried out in natural sea water, if the stock used is more than usually sensitive, the lethal effect of the highest copper concentrations possible in this medium is not as great as that of lower concentrations, an observation which agrees with the results obtained in artificial sea water and which perhaps lends further support to the thesis that the latter may be regarded as typical.

A number of experiments has been made in which cyprids of B. balanoides were exposed to mixtures of copper and mercury ; in some of these the concentration of copper was held constant and that of mercury varied, in others varying amounts of copper were added to constant amounts of mercury. In yet other experiments the cyprids were first exposed to one poison and then transferred to solutions containing the other.

These preliminary experiments show clearly that the effect of a mixture of these two poisons is complex, and that their combined effect is not simply additive, even if allowance is made for the greater toxic effect of mercury. Fig. 4 compares the effect of copper alone, mercury alone and of mixtures of the two poisons on the survival time, and Fig. 5 the effect of increasing concentrations of the mixed poisons for a constant exposure period of 6 hr., together with typical curves for the effect of each poison alone under similar conditions. Fig. 5 indicates that mixtures of copper and mercury are more toxic than either alone at lower concentrations, but that this relationship does not hold at higher concentrations.

Other experiments have been made using the cyprids of B. balanoides in an attempt to discover some difference in the mode of action or method of entry of copper and mercury. Though their results do not form part of the main theme of this paper, it is felt that they deserve mention as examples of the effect of other factors on heavy metal poisoning.

Cyprids of B. balanoides can survive for considerable periods in diluted sea water, and only small numbers (of the order of 20 %) are killed by exposure for 24 hr. to fresh water. If exposed to the action of copper or of mercury in diluted sea water, the toxicity of both these poisons is markedly reduced. Tables 4 and. 5 indicate the extent of this reduction.

These results also have a practical significance, since they must mean that copper and mercury leaching from an anti-fouling paint in estuarine conditions will be much less effective. As the leaching rate of these poisons is itself reduced in water less saline than sea water, it would seem that an anti-fouling composition, from which copper and mercury can be leached at an adequate rate in sea water, could be completely ineffective in the estuarine conditions prevailing in some harbours.

Hypertonic sea water (prepared by slow evaporation of natural sea water) also seems to have no effect on cyprids of B. balanoides, but the toxic effect of copper is markedly reduced in this medium. Some of the results are shown in Table 6.

As this reduction in toxicity, at least for copper, occurs both in hypo- and hypertonic sea water, it would seem unlikely that this phenomenon can be explained in terms of a direct osmotic effect, since that should mean that the larvae, if they were less sensitive in hypotonic, should be more sensitive in hypertonic solutions. A characteristic feature of the larvae immersed in hypo- and hypertonic sea water is their immobility—the antennae and thoracic appendages are withdrawn within the carapace and the larva rests on its side. Possibly this lack of movement is bound up with their decreased sensitivity to poisons.

It is possible that the rate of penetration of poisons is governed by the state of the bounding integument, and if this state is altered, e.g. by carrying out exposures to copper and to mercury in the presence of soaps, different rates of penetration and differences in toxic effect might occur. In carrying out experiments along these lines it was hoped that differences might be revealed in the rate of penetration of copper and of mercury. Unfortunately, as the results in Table 7 show, neither hope was fulfilled, the toxic effect did not vary with the soap concentration, and there was no significant difference between the effect of copper and of mercury with or without sodium oleate.

The two sets of estimations were not carried out on larvae of equivalent age, so that the copper/mercury ratio is not typical.

Two methods have been used to investigate the effect of exposure to copper and to mercury on the settlement of B. balanoides cyprids. At first, batches of cyprids were placed in low concentrations of the poisons in breffits, the total volume of solution was approximately 2 l., and observations on activity and the occurrence of settlement made at intervals. In this series of experiments the cyprids remained in the poison solution throughout. In a later series of experiments the period of exposure to the poison was limited and, after the end of the exposure period, the cyprids were washed with clean filtered sea water and then transferred to a further quantity of filtered sea water for observations of activity and settlement, which, as in the first set of experiments, were continued for some days. Each method has its disadvantages : in the first the poison concentration decreases over the period of observation (probably due to adsorption), and in the second the cyprids are finally transferred to a clean container which almost certainly does not present the optimum conditions for settlement.

In all cases, active cyprids were used from laboratory stocks in which appreciable settlement was taking place. If the cyprid of B. balanoides has to reach a certain physiological state before settlement is possible, it may therefore be assumed that a reasonable proportion of the stock used had reached that state.

The results of experiments carried out according to the first method, in which the larvae were exposed to low concentrations of poison over considerable periods, are set out in Tables 8 and 9.

The results of these experiments suggest that extremely low concentrations of both these poisons can prevent settlement, which agrees with the increasing sensitivity to copper during laboratory storage which was noted in the previous section (p. 280), and they also suggest that very low concentrations of copper are more effective in this respect than comparable concentrations of mercury. Though the latter observation is supported by observations on nauplii and free-swimming cyprids, where again the lowest concentrations of mercury used were less effective than the lowest concentrations of copper (though this relationship is reversed before concentrations are reached which are high enough to kill 50 % of the larvae), another explanation is possible. Concentrations of these poisons of the order used in these experiments are by no means stable, and the difference between the lowest copper and mercury solutions may merely be due to the more rapid disappearance (by adsorption on to the walls of the vessel or on to organic matter in suspension) of the latter element. Mercury certainly does disappear quickly as, in a later experiment on these lines, solutions which contained 0·019 mg./l. and 0·08 mg./l. at the beginning of the experiment contained no detectable mercury 4 days later.

Tables 10 and 11 give the results obtained using the second method, in which the period of poison exposure was limited and the further history of the larvae followed in fresh sea water.

These tables record estimates of the activity of the cyprids at the end of the poisoning period (XXXX, moderately active ; XXX, some active ; XX, a few active ; X, isolated individuals active ; O, no activity) and estimates of the amount of settlement (Good, Some, Light, Trace and No Settlement (NS)). A dash indicates no exposure for that period and concentration.

The results of short-period exposures to copper (Table 10) indicate that settlement is only possible after comparatively short periods of exposure to the lowest concentrations used and further that, though many or most of the cyprids may be active at the end of the exposure period, those individuals surviving in an active state are not capable of settlement. The results for mercury (Table 11) present rather a different picture. The concentrations of this poison used in these experiments had generally less effect on activity, but settlement subsequently took place mainly in stocks which had been exposed to higher concentrations of mercury.

The results for copper further confirm the inhibitory effect of very low concentrations of this poison on settlement, whereas those for mercury suggest something of a stimulatory effect for the higher concentrations used. The results of neither set of experiments can be regarded as wholly satisfactory, and it would be unwise to speculate further on results obtained under such artificial conditions.

The experiments on the sensitivity of the cyprid of B. balanoides during storage in the laboratory, and on the effect of copper on settlement, suggest that this larva becomes very sensitive to copper at the time of attachment. A series of tests was therefore made on cyprids of this species which had settled and which were metamorphosing into the adult form. The morphological changes which occur during this process were not followed in detail, but it was found possible to recognize five stages in the sequence (Fig. 6) which may be described as follows.

Stage A. Immediately after attachment; the body of the cyprid projects upwards from, and makes an acute angle with, the substratum. The antennae are firmly anchored and permanent attachment has been achieved. This stage is probably short-lived, but it lasts long enough to be recognized in an appreciable proportion of a population of recently attached cyprids.

Stage B. The body of the cyprid now lies close to the substratum, the anterior end is appreciably flattened and the tissues internally show a distinct median grove anteriorly.

Stage C. The major external changes are now complete, and the form of the adult has been assumed. Shedding of the cyprid integument is not complete, as the thoracic appendages still remain within their original integument. The cyprid carapace, therefore, remains attached to the young barnacle and seems to remain so for some little time.

Stage D. The cyprid carapace has now been shed completely. Calcification has not yet begun and the perimeter of the basis bears scattered tufts of fine setae.

Stage E. Calcification has begun and the outlines of the primary compartments are clear. The paired eyes, clearly visible until this stage, are no longer apparent.

Small populations of B. balanoides cyprids which had settled, mainly on Mytilus shells, during the 24 hr. immediately preceding the experiment, were placed in a series of copper solutions and observations made on their further progress. All the individuals used were initially in Stage B.

The results of two experiments are shown in Table 12.

The results given in this table show clearly that the sensitivity during metamorphosis is far lower than either that of the cyprid just before settlement or (as will be shown in the following section) of the barnacle after metamorphosis. Metamorphosis can continue in all the concentrations of copper possible in natural sea water, though it should be noted that, as calcification only begins when individuals which have metamorphosed in the two lowest concentrations used are transferred to fresh sea water, death occurs in the remaining cases immediately after this process has been completed.

The low sensitivity of the cyprid during metamorphosis is probably the explanation of a phenomenon frequently noticed in raft exposures of anti-fouling compositions—that settlement and metamorphosis can take place on a surface too toxic for the continued life of the barnacle after metamorphosis. Individuals settling under such circumstances do not become calcified and die within a few days after metamorphosis.

Experiments have also been carried out on the sensitivity of the cyprid of Balanus crenatus during metamorphosis. The main external changes during the metamorphosis of this species are generally the same as those of B. balanoides, with two exceptions :

• Stage C is virtually absent, as the cyprid carapace is rapidly shed at the conclusion of metamorphosis. (This stage is therefore omitted from the tables below.)

• Calcification does not begin so soon after the completion of metamorphosis. Further progress is, however, marked by the appearance of the compartments and of the outlines of the scuta and terga. The stage shown as D’ in the tables below refers to individuals which have reached this condition.

The results of these experiments are summarized in Tables 13 and 14. All the cyprids used were initially in Stage B.

These results show clearly that the cyprids of B. crenatus are able to complete their metamorphosis in all the concentrations of copper used but, judging by their failure to progress beyond Stage D (except for the controls), they can be killed by low concentrations of copper immediately metamorphosis has been completed.

The measurements of the copper content of the solutions (and for the mercury solutions given in Table 14) were made at the end of the exposure period. It is most unlikely that the initial concentrations of the poison solutions were below the nominal values indicated, so that the difference between the minimal and measured values gives a rough guide to the range of concentrations existing in these solutions during the test period.

The results with mercury are not so clearly defined as those obtained using copper. Some metamorphosis appears to have been possible in all the concentrations of mercury used, but its extent bears little relationship to the concentration. Further, if the state after 10 hr. exposure is compared with that after 24 hr. exposure it is seen to be the same, with the exception of the 0·5 and 1·0 mg./l. solutions, in both cases. A possible interpretation of these results is therefore that, above a concentration of 1·0mg./l. (or less), mercury is toxic, but that it takes a little time to exert its effects. Thus, in a population of metamorphosing cyprids, in which some have recently entered Stage B and some are much further advanced, the latter may be able to complete their metamorphosis before sufficient poison has been absorbed to stop the process.

Thus it would appear that the cyprid of B. crenatus is as insensitive to copper during metamorphosis as that of B. balanoides, and it may be suggested that the former species is much more sensitive to mercury. It is of interest to note that the sensitivities of the metamorphosing cyprid of B. crenatus are similar to those of the free-swimming larva, whereas the sensitivity of the metamorphosing cyprid of B. balanoides is very different (at least for copper) from that of the free-swimming larva immediately before settlement.

This series of studies on the sensitivity of the larval stages of B. balanoides and B. crenatus was completed by a series of estimations of the sensitivity of barnacles after metamorphosis.

For this purpose, individuals which had settled on ground glass microscope slides were used. Before each experiment, the slides used were cleaned as thoroughly as possible (to remove as much organic matter as was practicable, so as to present the minimal surface for adsorption of the heavy metal ions) and then placed in the poison solutions. Each container had a capacity of approximately 175 ml., and a slow circulation (4–5 I./24 hr.) of the poison solution was maintained through the containers for the whole of the exposure period. Normally the latter was 6 hr. (so that the results of these experiments are comparable in this respect with those on free-swimming cyprids) but in a few experiments it was 24 hr. in length.

Immediately after the end of the experiment, the slides were removed from their containers, washed in running sea water and placed under sea-water circulation. Each was then examined and the activity of each individual recorded according to the following scheme:

Full and rapid movements of the cirri possible.

Some cirral movement possible, but cirri never fully extended, frequency of movement may approach normal.

If the opercula are closed, they respond to touch; if the.opercula are open and the cirri are protruding the latter can execute slow (often spasmodic) movements or can be stimulated to do so with a needle.

No response of closed opercula to touch and no response of cirri, if protruding.

The slides were then reimmersed in the sea and again examined a week or 10 days later. After this interval, those which had been killed during the exposure to poison or which had died later were immediately obvious, as their shells were empty, whereas those which had survived or recovered were fully active.

Tables 15 and 16 show the results of two typical experiments in which B. crenatus was exposed, in one case to copper and in the other to mercury.

These tables not only illustrate the way in which the results of these tests were assessed, they also show something of the characteristics of copper and mercury poisoning of barnacles. It will be noted that the ‘Apparent Survival’ is a more reliable guide to ultimate survival when copper has been used as a poison than is the case when the barnacles have been exposed to mercury, which suggests that copper is more immediate (and possibly less permanent) in its effects than mercury.

Comparisons of the sensitivity of B. balanoides and B. crenatus to these two poisons reveal remarkable differences. Table 17 shows the results of determinations made on these two species over the period following metamorphosis. Those on the left-hand side of the table for each species are the results obtained for the youngest specimens tested; those on the right-hand side of the table those for the oldest specimens tested.

The experiments on B. balanoides cover a period of about $212$ months from settlement, those on B. crenatus a period of just over 3 months from settlement.

Copper and mercury seem roughly equi-toxic (weight for weight) to B. balanoides’, B. crenatus is rather more sensitive to copper but very much less sensitive to mercury. Further, the sensitivity of B. balanoides remains much the same over the range of ages examined—in particular, the youngest specimens have much the same sensitivity as the oldest, whereas the youngest specimens of B. crenatus are distinctly less sensitive to both poisons than older individuals, though the ratio between the two poisons remains roughly the same.

Extension of the exposure period to 24 hr., though it naturally affects the median lethal concentration, does not cause any marked differences in the relative effect of the two poisons. For this exposure period the median lethal concentration for copper for B. balanoides is 0·32 mg./l., and that for mercury 0·36 mg./l. The corresponding values for B. crenatus are 0·19 mg. Cu/1. and 1·35 mg. Hg/1.

The results described in this paper reveal considerable differences in sensitivity to copper and to mercury between different stages in the life history of one species and between corresponding life-history stages of the two species.

Thus, the naupliar stages of B. balanoides are highly sensitive to copper, the cyprid, at least immediately after it is formed from the sixth stage nauplius, is much less sensitive to this element. During the life period of the cyprid the sensitivity again increases, only to decrease markedly once settlement has occurred and metamorphosis begun; when metamorphosis is completed the sensitivity again increases. The sequence of changes for B. crenatus follows much the same general course, except that the cyprid of this species is so insensitive to copper that any differences in sensitivity between the free-swimming and the metamorphosing larva cannot be detected using natural sea water, and the sensitivity relationships of the adult B. crenatus differ from those of B. balanoides. Such differences emphasize the fact that the results which are obtained with one species cannot be applied without reservation to another, even if the two species are closely related.

The causes of these differences in sensitivity between stages and species are at present obscure. This obscurity is largely caused by the lack of knowledge of the effects of heavy metal poisons, and partly by marked differences in morphology and physiology between the various stages in the life history of a barnacle. There appear to be broadly two schools of thought on the mode of action of heavy metal poisons : one, that these elements never penetrate the tissues, but exert their effect by precipitating proteins on the surface of the animal and so interfering with such vital processes as respiratory exchange (Carpenter, 1927, 1930; Jones, 1947); the other, that heavy metals penetrate the tissues, where they act intracellularly as enzyme poisoners or protein precipitants. There is evidence (Clarke, 1947) which indicates that copper is taken up into the tissues of barnacles, and Mr W. R. Hunter (personal communication) has recently demonstrated that small amounts of copper cause a marked decrease in the oxygen uptake of Marinogammarus marinus, whereas small amounts of mercury have no significant effect on this process. This suggests that copper penetrates the tissues and has an effect on intracellular metabolic processes, particularly, it might be suggested, on the respiratory enzyme system. Though it is dangerous, without direct experimental proof, to suppose that these results obtained for M. marinus apply also to barnacles, it is reasonable to look for similar effects in the latter. In this connexion, the marked decrease in sensitivity to copper during metamorphosis is of particular interest since, if copper has an effect on the respiratory enzyme system and if respiration were anaerobic during metamorphosis, an explanation could be advanced for the decrease in copper sensitivity.

The effect of increasing copper concentrations in artificial sea water is a further point of interest. These effects may be due to the medium employed, but there are indications from the unpublished work just quoted that they are more likely to be due to the effects of the poison used. If, in these experiments, copper is acting as a direct poison and also affecting some metabolic process which in some way controls the uptake of the poison, it is possible to interpret the results rather more fully. As has been mentioned earlier, copper may not only interfere with its own action, it may also affect the action of mercury when the two poisons are present together.

Measurement of the sensitivity to copper may also be a valid means of assessing the physiological state of a free-swimming cyprid, and this may prove a useful means of investigating the effect of variations in environmental factors on this stage. In particular, experiments have been begun on the effect of drying on the cyprid of Balanus balanoides, and the preliminary results of these experiments suggest that such studies may provide an explanation of the intertidal distribution of the adults of this species.

There is no reason to assume that variations in effect on intracellular processes are the sole causes of the differences in sensitivity recorded in this series of experiments. Penetration of the poison may be effected much more readily in one stage than in another. For example, in the naupliar stages an open gut is present, whereas in the cyprid the gut is closed. Penetration of poison may therefore take place more readily in the former than in the latter (Waterhouse, 1946, has demonstrated the presence of copper in the walls of the gut of Lucilia cuprina when fed on cultures containing this element), and the difference in sensitivity between the nauplii and the cyprids of both species investigated may be due largely to this cause.

Finally, apart from biological factors which make interpretation difficult, it is evident (Barnes & Stanbury, 1948) that the solution of either copper or mercury salts in sea water is a complicated process, and the existence of a number of complexes, the proportions of which may vary at different concentrations, adds another, and possibly an overriding factor to a situation already sufficiently complex if considered only in its biological aspects.

It is evident that too little is known, both of the effects and of the conditions of penetration of these poisons, fully to interpret the results described in this paper. More recent work, as yet unpublished, will take the interpretation a stage further.

The authors wish to record their thanks to the Marine Corrosion Sub-Committee for their permission to publish this work, and they also wish to acknowledge the help and encouragement they have received from Prof. J. E. Harris. The senior author is also much indebted to his colleagues, Mr M. W. H. Bishop and Mr W. R. Hunter for their help in the later stages of this work.