i65
THE TOXIC ACTION OF HEAVY METAL SALTS
ON THE THREE-SPINED STICKLEBACK
(GASTROSTEUS ACULEATUS)
BY J. R. ERICHSEN JONES.
(Department of Zoology, University College, Wales, Aberystwyth.)
(Received 24th November, 1934.)
(With Seven Text-figures.)
A. INTRODUCTION.
an aquatic animal is placed in an aqueous solution of a toxic metallic salt of
sufficient concentration, it dies in a time which is called the survival time. This
time bears a relation to the concentration which may be expressed graphically by
plotting experimentally determined survival times against the molar concentration
or normality, the interpolated graph being called the survival curve.
It was shown by Powers (1917) that in the case of the goldfish (Carcassius
carcassius), each salt has a more or less definite lower limit of concentration below
which the survival time is indefinite and protracted. This concentration was termed
by Powers the threshold concentration.
Metallic salts can be divided into two groups according to whether this threshold
concentration is high or low. In the first group may be placed the majority of the
salts of the metals of the alkalis and alkaline earths which are relatively harmless
to fresh-water fish at concentrations below o-ioN. The following are some threshold
concentration values for such salts given by Powers:
WHEN
Lithium chloride
Sodium chloride
Magnesium nitrate
Strontium nitrate
Calcium nitrate
0
...
...
...
...
...
0-125^.
o-2$N.
o-1 jN.
0-155^.
O-IJN.
These are all for Carcassius at 18 C.
In the second group are the salts of the heavy metals—lead, copper, zinc, nickel,
cadmium and mercury, which are definitely toxic at concentrations from o-ooooiiV
upwards.
The toxic action of lead nitrate and other heavy metal salts on the minnow
(Leuciscus phoxinus) has been investigated by Carpenter (1927, 1930), who showed
that lead nitrate solutions were definitely lethal to this fish down to a concentration
of approximately o-ooooc^iV.
166
J. R. ERICHSEN JONES
Furthermore, it was shown that the death of the fish was due to the formation
of a film of coagulated mucus over the gills and body surface as a result of chemical
action between the lead ion and the mucus secreted by the gills. This film so impeded
respiration that the fish died from suffocation.
This conclusion was supported by the following evidence:
(1) When in the lead solution the fish displayed great respiratory distress, the
opercular movements were more rapid, but the rate of evolution of carbon dioxide
was less than that of controls in tap water.
(2) After death the film of coagulated mucus on the gills was clearly visible,
and treatment with dilute ammonium sulphide caused the film to turn black,
indicating the presence of lead.
(3) If the fish were removed from the solution a reasonable time before death,
and placed in an abundant supply of clean water, or in a running supply, the mucus
film was shed off and the fish completely recovered, little the worse for its experience.
(4) All the lead removed from solution could be recovered from the gills by
treatment with dilute acetic acid; chemical analysis showed the rest of the body to
be perfectly free from lead.
Carpenter also showed that the toxic action of other heavy metal salts, cadmium
sulphate, ferrous sulphate, ferric chloride, copper sulphate, copper chloride and
mercuric chloride, was exactly similar in nature; death was always due to impediment of respiration as a result of the metallic ion's reaction with the mucus to form
an impermeable film.
B. EXPERIMENTS WITH THE STICKLEBACK
(GASTROSTEUS ACULEATUS).
The writer has studied the toxic action of solutions of salts of lead, copper, zinc,
nickel and cadmium on the common three-spined stickleback (Gastrosteus aculeatus),
and the observations and conclusions of Carpenter have been found to apply to
this species of fish also. In the solutions the sticklebacks always exhibited the same
symptoms of respiratory distress noted by Carpenter, their behaviour just before
death was similar and after death the gills were seen to be coated with a film of
coagulated mucus which reacted with ammonium sulphide to give black precipitates
except in the case of cadmium when the precipitate was yellow. It was therefore
concluded that the physiological effect of these salts was the same in the case of the
stickleback as in that of the minnow.
Solutions of sodium nitrate, sodium chloride and sodium sulphate are relatively
harmless to sticklebacks even at concentrations above o-i6N. Sticklebacks will
live for days in a o-2oN solution of sodium chloride and will live over 6 hours in
a crzoN solution of sodium nitrate or sodium sulphate.
In o-2oN solutions of copper sulphate, copper nitrate and copper chloride on
the other hand their survival time is a matter of minutes. In O-O2N solutions of
copper sulphate, nitrate or chloride they survive barely an hour, whereas in O-O2N
solutions of the corresponding sodium salts they live for days or even weeks.
Toxic Action of Heavy Metal Salts on the Three-Spined Stickleback 167
This seems to indicate that the anions NOa~, Cl~ and SO<~ ~ are relatively harmless at concentrations from o-zoN down, and that the high toxicity of the copper
salts is due mainly to the activity of the cation. This also would apply to the nitrates
and sulphates of zinc and nickel and cadmium.
This being so, it might be expected that copper nitrate and copper sulphate in
equimolar concentration should have the same or approximately the same toxicity,
the concentration of copper ion being the same in both cases. This, however, is not
the case, as at all concentrations copper sulphate is markedly less toxic than equimolar copper nitrate. Similarly zinc sulphate is much less toxic at the same molar
300
K—x— Zinc sulphate
Zinc nitrate
•3 200
.9
-£
HO
100
70
JO
30
20
in
0
0-06
OI0 0-12 0-M
0-20 0-22
0-30
Normality
Fig. 1. Survival curves for Gattrotteus in zinc nitrate and zinc sulphate.
concentration than zinc nitrate and the same holds for the nitrates and sulphates
of cadmium and nickel.
The results obtained with zinc will be considered first. Fig. 1 givea the survival
curves for zinc nitrate and zinc sulphate at 170 C.
Each plotted point represents the mean survival time of two sticklebacks in
200 c.c. of solution. Ordinary distilled water (not glass distilled) was used for
making up the solutions, and the solutions were used in wide-mouthed glass jars
which were stood in a shallow tank through which a steady stream of tap water
flowed round the outside of the jars. This maintained the temperature of the
solutions at 17° C. Tap water from the same source flowed through the tank in
which the stock of fish was kept, so the fish suffered no temperature change on
168
J.
R.
ERICHSEN
JONES
transference to the solutions. The survival time in all cases is the experimentally
observed time between the introduction of the fish and the final cessation of the
opercular movements. Towards the end of the survival time the fish were stimulated
by touching with a glass rod. When a fish failed to respond to such stimulation by
breathing or swimming it was taken to be dead and the survival time noted.
The distilled water used was thoroughly aerated before making up the solutions.
Sticklebacks used as controls survived 2-3 days in this distilled water. Controls
live without food for a week or more in tap water or rain water. Unfortunately
tap water could not be used for making up the solutions as the high carbonate
content of the Cambridge tap water precipitated much of the copper, zinc, etc. as
carbonates, and rain water was not available in sufficient quantity.
The early death of the controls in the distilled water was assumed to be due to
the presence of traces of copper from the still. This copper content was probably
always very small compared with the quantity of toxic ion added, so little error
resulted from its presence. In fact, a few experiments showed that the survival
times of sticklebacks in solutions made up with rain water were no longer than
those treated with solutions made from the distilled water.
At low concentrations some of the survival time determinations were repeated
with a much greater solution volume to make sure that the absolute content of salt
was sufficient. It was found that down to a concentration of o-ooiN, increasing
the solution volume had no appreciable tendency to diminish the survival time and
200 c.c. was therefore considered adequate volume down to this concentration
which was the lowest employed in the case of zinc. This precaution was taken, as
it was shown by Carpenter that, in the case of the minnow, at low concentrations
the toxic salt is so rapidly precipitated by the mucus secreted that the concentration
falls appreciably during the survival time and the latter is unduly prolonged.
The writer has found that at any concentration the survival time is considerably
influenced by the sex and size of the fish, the season of the year, the locality from
which the fish were obtained, the time they had been in captivity, etc.; the individual
variation in survival time which resulted from varying these conditions was surprisingly great, often amounting to 70 per cent.
In order to obtain consistent results and an accurate comparison between the
two salts the following precautions were taken:
(1) All the fish used were mature males and measured 42-50 mm. in length
(tip of mount to end of tail fin).
(2) All were obtained from the same stream on the same day.
(3) The experiments for zinc nitrate and zinc sulphate were carried out concurrently and under identical conditions. A sufficient number of fish was collected
and a number of sulphate and nitrate experiments performed each day. The whole
series of experiments occupied four consecutive days.
The survival curves in Fig. 1 which are interpolated freehand, show clearly
that at all concentrations the nitrate is definitely more toxic than the sulphate. The
possible reasons for this are now to be considered.
Toxic Action of Heavy Metal Salts on the Three-Spirted Stickleback 169
There is no appreciable difference between the pR values of the two series of
solutions. The/>H of the zinc nitrate solutions was found to vary from 5-5 at o-zoN
to 6-8 at o-ooiiV. That of the sulphate solutions showed an almost exactly similar
variation.
Fresh water ofpH 5-5, that of the most acid of the solutions, is practically harmless to sticklebacks which live over a week in fresh water maintained at this/>H value.
Therefore it seems safe to conclude that the acidity of the solutions cannot
account for the difference of toxicity of the two salts.
0
0-05
01
0-2
0-3
0-4
0-5
Normality
Fig. 2. Conductance ratio curves for zinc nitrate and xinc sulphate.
As the physiological action of the salts takes place externally to the body it
seems unlikely that the difference results from difference of rate of permeability.
Similarly it does not seem at all likely that difference of osmotic pressure is the
cause, especially as the effect is well marked at concentrations far below isotoxicity.
However, there is a most marked relation between the relative toxicity of the
two salts and their relative electrical conductivity. The theory of electrolytic
dissociation is discussed at length in all text-books of physical chemistry and will
only be outlined here.
It was believed by Arrhenius that for all electrolytes the fraction A/A<,, or the
equivalent conductivity at any particular concentration divided by the equivalent
J. R. ERICHSEN JONES
170
conductivity at infinite dilution, represented the fraction of the salt dissociated
into ions at that concentration.
Arrhenius believed that the mobility of the ions, that is their speed of movement
under the potential gradient, was the same for all concentrations of any one salt at
a fixed temperature, for instance the copper ion in a o-ioiV solution of copper
sulphate was supposed to travel at the same rate as the copper ion in a o-ooiN
solution of copper nitrate when subjected to the same potential gradient at the
same temperature. The fall in equivalent conductivity which was observed to take place with
310
rise in concentration was believed to result from
diminution in degree of dissociation, there being
relatively fewer ions to carry the current.
More recent work on the electrochemistry of
solutions appears to show that for strong electrolytes the fall in conductivity observed on rise in s
- Zinc sulphate
concentration is due, not to fall in degree of
Zinc nitrate
ionisation but to decrease in the mobility of both
ions. In other words, at high concentrations
strong electrolytes are dissociated to practically
the same extent as at infinite dilution, but as a
result of interionic attraction and other factors
the speed of movement of the ions is diminished
and the electrical charge carried in a given time
under the same potential gradient is less, i.e. the
equivalent conductivity decreases.
The factors which influence the mobility of
the ions also seem to govern their chemical
activity and for strong electrolytes it is found
that the conductance ratio (A/A,,) can be taken as
.
.
.
...
a- .
Fig. 3. Survival curves for Gaitrosteui in
an approximate value for the activity coefficient.
l i n c n i t r a t e md zinc BU i phate .
In other words at any concentration the absolute
chemical activity of a strong electrolyte is proportional to the product of its concentration in normality and its conductance ratio at that normality—(A/A,,) N.
The conductance ratio for zinc sulphate falls much more rapidly as the concentration increases than does that of zinc nitrate. This is shown in Fig. 2. Normality
is plotted as abscissa and conductance ratio as ordinate. These curves were interpolated from the results of large numbers of experimental determinations of the
equivalent conductivity of zinc sulphate and zinc nitrate solutions at 170 C. The
usual arrangement of Wheatstone Bridge, induction coil, telephones and platinised
electrode cell was used in all cases.
Thus at all concentrations the chemical activity of zinc sulphate appears to be
decidedly less than that of zinc nitrate of equimolar concentration. In fact, if the
survival curves for zinc sulphate and zinc nitrate are plotted, not against normality
but against (A/Aj) iV, the two curves become almost coincident. This is shown in
Toxic Action of Heavy Metal Salts on the Three-Spined Stickleback 171
Fig. 3. Here the survival times are the same as those in Fig. 1, but instead of concentration (A//\,) N is plotted as abscissa. That is, for both salts each plotted point
gives the survival time at the normality, which, multiplied by the conductance ratio
at that normality, gives the value denoted by the corresponding abscissa.
The results show that the difference in toxicity between the two salts results
from differences in chemical activity, and that for both salts the chemical activity
—or toxicity—is proportional to the product of the concentration and conductance
ratio.
140
-*-Copper sulphate
—O-®—Copper nitrate
•A'-'Copper chloride
3
's
.S
X Copper sulphate
O Copper nitrate
A. Copper chloride
i
IOO<
B
I
•a
1
0
0-02 (XM
007
010 "
0-14
02
0-25
0-3
70
0
002
0-04
0-07
0-10
0-14
Normality
Fig. 4. Survival curves for Gastroitetti in copper salts.
Fig. 5. Survival curves for Gastrosteus
in copper salt*.
Attempts to establish the same relationship between lead salts were not successful ; a survival curve for lead nitrate was obtained and was found to be very similar
to that obtained by Carpenter for the minnow. However, lead acetate solutions
were found to decompose so rapidly when fish were placed in them that satisfactory
survival time experiments were impossible. It was noted that lead acetate was less
toxic at the same molar concentration as lead nitrate.
Lead sulphate is not sufficiently soluble for its comparative toxicity to be
estimated at concentrations where its equivalent conductivity differs materially
from that of lead nitrate.
iy 2
J. R- ERICHSEN JONES
Difficulties were also encountered with mercury salts. Mercuric chloride is not
an electrolyte and mercuric nitrate decomposes rapidly in solution. A few experiments were made with cadmium nitrate and cadmium sulphate and the results were
similar to those for zinc.
Experiments were performed with three cupric salts, copper nitrate (Cu(NOs),),
copper chloride (CuCl,) and copper sulphate (CuSO*). The methods employed
were similar to those employed for zinc, and exactly the same precautions were
taken to ensure consistency of results and accuracy of toxicity comparison. For
3001
JIO
(.- Nickel sulphate
•— Nickel nitrate
3
C
' g 200
200
—*—*•- Nickel sulphate
—©-©- Nickel nitrate
.S
t
en
100
60
60
40
40
*~Z"
20
20
10
" 0 CHH 004
100
0H0
0-U
0-20
0 0-02 OKM 0 06
010
0-14
Normality
Fig. 6. Survival curves for Gastrosteus
in nickel saltB.
Fig. 7. Survival curves for Gastrojtetis
in nickel salts.
each pair of sticklebacks 200 c.c. of solution was used down to a concentration of
o-ooiN and was found to be adequate volume at this concentration. At concentrations below this value, the solution volume was increased in proportion to keep
the absolute salt content the same, i.e. at C-00002./V, the volume was 10,000 c.c.
This was the lowest concentration used.
Values for the conductance ratio for the three salts were determined as in the
case of zinc, and it was observed that that of the nitrate and chloride dropped slowly
with rise in concentration while that of the sulphate dropped more rapidly. The
survival curves for the three salts plotted against normality and (A/A,,) N respectively
are given in Figs. 4 and 5. It will be seen that the three curves in Fig. 5 are
Toxic Action of Heavy Metal Salts on the Three-Spined Stickleback 173
remarkably coincident, i.e. the toxicity seems to be determined by the normality
multiplied by the conductance ratio.
A very similar result was observed with nickel salts. Survival curves were
plotted for nickel sulphate and nickel nitrate against normality and (A/A,,) N. The
results are given in Figs. 6 and 7.
The agreement between the curves in Fig. 7 is not as good as in those for zinc,
but it is obvious that the toxicity is determined rather by (A/A,,) N than by the
normality alone.
The nickel experiments were conducted in an exactly similar manner to those
performed with zinc salts.
Thus for the stickleback the toxicity of all the stable and soluble heavy metal
salts appears to be determined by (A/A,,) N. The same relation has been found to hold
in the case of other aquatic animals, namely Polycelis nigra and Gammarus pulex,
as will be shown subsequently.
SUMMARY AND CONCLUSIONS.
1. The observations and conclusions of Carpenter (1927, 1930) regarding the
toxic effect of heavy metal salts on the minnow (Leticiscus phoxinus) are shown to
apply equally well to the three-spined stickleback (Gastrosteus aadeatus).
2. It is shown that in the case of heavy metal salts the toxic effect of hypertonic
solutions is due chiefly to the metallic cation, the toxicity of the anion being relatively
small, while in the case of hypotonic solutions the toxicity is due entirely or almost
entirely to the cation.
3. Nevertheless, in equimolar concentrations, different salts of the same metal
do not have equal or approximately equal toxicity. At the same molar concentration
the sulphates of heavy metals are much less toxic than the nitrates or chlorides.
This is not due to difference of/>H.
4. This difference in toxicity is shown to be closely related to difference in
relative electrical conductivity; the toxicity of nitrate, chloride and sulphate is
determined, not by the normality alone, but by the product of the normality and
conductance ratio.
5. It is concluded that the factors responsible for the lower electrical conductivity
of the sulphate also lower the chemical activity of the sulphate so that the toxicity
falls in proportion to the conductivity.
The work was all carried out at the Laboratory of Experimental Zoology,
University of Cambridge, under the supervision of Dr J. Gray.
REFERENCES.
CARPENTER, K. E. (1927). Brit. J. exp. Biol. 4, No. 4.
(1930). J. exp. Zool. 7, No. 4.
POWERS, E. B. (1917). Illinois biol. Monogr. 4, No. 2, pp. 1-73.