1. In 279 cases 97 % agreement was obtained between Zondek-Aschheim and Xenopus laevis pregnancy tests.

  2. Feeding once a week with either beef or liver is sufficient to maintain ovarian growth in Xenopus laevis but a diet of liver is more propitious if the frequency of feeding is less. The superiority of a liver diet is not due to its greater manganese or vitamin content.

  3. The effect of overcrowding upon the reproductive activity of Xenopus laevis is not due to pollution alone but also to some form of interaction between the animals themselves. The maximum density level compatible with ovarian growth is higher if the water is kept running than if it is only changed after feeding.

  4. Ovulation in Xenopus laevis as a response to injection of tissue extracts in doses up to 5 g. of tissue is specific to water extracts of the anterior lobe of the pituitary.

  5. Instructions for the technique of the Xenopus pregnancy test are outlined.

The control of the amphibian sexual cycle has been demonstrated by experiments on the removal of the pituitary and subsequent injection of pituitary extracts in Xenopus laevis by Hogben and his collaborators (Hogben, 1930; Hogben et al. 1931 ; Bellerby & Hogben, 1938). The earlier work on the female cycle was carried out in Cape Town where preliminary experiments on the urine of pregnant women were undertaken. Bellerby (1933), who carried out further work in Prof. Hogben’s department in London, showed that Xenopus can be transported from South Africa to Europe and kept in suitable condition for use as a test animal for pituitary extracts. This depends on the fact that Xenopus does not extrude all its ripe eggs simultaneously as do Bufo and Rana and it never ovulates spontaneously under normal laboratory conditions. The extrusion of the fertilizable eggs which lack the normal gelatinous envelope is a highly specific response to gonadotropic substances. Since its recognition does not involve killing the test animal, which can be used repeatedly for the same purpose, Xenopus has a very great advantage over mammalian material. An additional minor advantage is the fact that weekly feeding suffices to maintain the animal in a satisfactory condition. Between 1 and 6 months after conception 2 c.c. of untreated urine directly injected into the dorsal lymph sac induces ovulation within 18 hr. of clawed toads kept at 22° C.

Shapiro & Zwarenstein (1933, 1934, 1935) and Bellerby (1934) have now independently obtained sufficiently extensive records to justify the claim that ovulation of Xenopus can also be used as a test for the gonadotropic substance in pregnancy urine. The writer has recently collaborated with Prof. Crew’s Pregnancy Diagnosis Laboratory in a parallel study on the results of the Zondek-Aschheim and of the Xenopus test. Results of these observations will be published more fully at a later date. A comparison of 279 cases tested by both methods yields the following figures (Table I).

Table I
graphic
graphic

There were in fact six discrepancies, five specimens gave negative Xenopus and positive Zondek-Aschheim results. Four of these five were done in early pregnancy, and further concentration of the urine might have given more parallel results. One case gave a positive Xenopus and a negative Zondek-Aschheim result, but the latter test is not considered infallible by Robson (1934) who claims 1-3 % inaccuracy for it.

It has been shown by Orent & McCollum (1931) that addition of traces of manganese to the normal diet of male rats enormously increases the growth of the gonads, while a manganese-free diet results in atrophy.

In captivity Xenopus laevis has generally been fed on beef which has a very low manganese content (about o-oi mg.%). It was decided to investigate the effect of feeding the toads on liver, which has a manganese content of about 0·25 mg. %, and also to add traces of manganese to both beef and liver diets.

The determinations of manganese contents were made with a Hilger quartz spectrograph purchased with a grant made to Prof. Hogben’s department by the Halley Stewart Trust. The tissues were extracted by boiling with perchloric acid. The boiling was continued until all the perchloric acid was driven off. The residual liquor was burned on a filter paper spill in an oxy-coal gas flame in front of the spectrograph. Standards were superimposed on the same spectrum by moving the plate holder laterally about 1 mm.

Beef contains only a trace of vitamin E and an attempt was made to improve the sensitivity of the ovary to injections of A.L.P. by addition of vitamin E to the diet. The vitamin E used was supplied by the Glaxo Research Laboratories in the form of wheat-germ oil five times concentrated and dissolved in cod-liver oil. A batch of control animals was given cod-liver oil alone in addition to their diet.

Two sets of seventy-two female toads were used. One set (A) had been fed once a week and the other set (B) had been starved for 6 weeks prior to the experiment. Each set was divided into six batches of twelve animals and kept in 201. of running water at 20° C. Each batch was fed once in 3 weeks according to Table II, and the tanks were cleaned out 24 hr. after feeding. The duration of the experiment was 5 months. Each batch was weighed before and after feeding. The food intake was found by subtraction.

Table II
graphic
graphic

100 g. minced beef and 100 g. of chopped liver was supplied at each feeding. Batches 4 and 6 were injected with 1 c.c. of 0·01 % MnCl2, solution into the dorsal lymph sac, 1 week before each feeding.

Half a c.c. of the cod-liver oil or of the wheat-germ solution was given by hand just before each feeding. The beef which followed kept the oil in the stomach.

At the end of the experiment the toads were injected with an extract prepared from dried anterior lobe powder. Each toad was given the equivalent of 1 g. of fresh tissue and the number ovulating in each batch was recorded. While the results of this experiment, as shown in Table II, show that liver is a better diet than beef, they also show that addition of neither manganese, nor vitamin E, nor cod-liver oil, to a diet appears to have any appreciable effect on the animals. Hence the apparent superiority of a liver diet is not due to its greater manganese or vitamin E content.

A further experiment was carried out to see if this is true when the animals are fed once a week. Two batches of thirty toads were fed once a week on beef and on liver respectively. Each batch was kept in 10 1. of water changed once a week 24 hr. after feeding. Over a period of 4 months the growth curves of the two batches were almost identical. The temperature was 15 ± 1° C. The total food intake figures were 1497 for the beef-fed and 1676 for the liver-fed. At the end of the experiment the toads were injected with an extract of the anterior lobe of the pituitary. The dose given was half as large as in the preceding test, being the quantity requisite to evoke response in 50 % animals of their body-weight class when fully fed. The test showed that sixteen of the thirty beef-fed and seventeen of the liver-fed responded.

That is to say both sets gave the 50% response appropriate. The result of the preceding experiments indicates that feeding once a week with either beef or liver is sufficient to maintain ovarian growth, under stated conditions of temperature and water volume, and that a diet of liver is more propitious if the frequency of feeding is less.

It has been shown by Alexander & Bellerby (1938) that overcrowding of Xenopus laevis leads to retrogression of the ovaries and to diminution of the food intake, even though abundant food is available. They found that 40 g. animals required a maximum density of 3·3 toads per litre if they were to give an 80 % response to injection. These authors were unable to explain this result. The water of their experimental animals was changed only once a week. In so far as there is a specific overcrowding effect which is not due to insufficient food resulting from competition there are several possibilities to explore. One is that it is due to diminished dermal respiration due to low oxygen content of the water. Another is that it is a toxic effect due to accumulation of materials secreted by the skin or excreted with the urine and faeces of the animals. Both of these possibilities can be investigated by a type of experiment in which there is a pollution gradient, illustrated in Fig. 1.

The five aquaria were of the same dimensions and housed twenty-five toads. The outlet holes were bored in such a position that each tank contained 5 l. of water.

The inlet was adjusted to allow 5 1. of water at room temperature to flow in 6 hr. Each aquarium was independently aerated.

The animals were fed once in 2 weeks over a period of 3 months, each tank receiving 100 g. of minced beef. During feeding the flow was stopped and 24 hr. after feeding the water was changed and the flow restarted. The twenty-five toads from each tank were weighed in bulk before and after feeding and the food intake figure arrived at by subtraction. Determinations of dissolved oxygen were made at regular intervals. The results (Table III A) show that pollution is a major cause of diminished food intake found in overcrowding and that oxygen content of the water does not greatly affect the result.

Table IIIA
graphic
graphic

It does not follow from this that pollution is of itself sufficient to account for the overcrowding effect. An alternative procedure showed conclusively that neither oxygen deficiency, nor available food, nor pollution, provide a sufficient explanation of this phenomenon, and there is a residual overcrowding effect when all these three factors are excluded. This was shown by keeping toads in an aquarium partitioned exactly halfway with wire netting fixed vertically. Ten toads (batch A) were placed on one side of the partition and thirty toads (batch B) on the other side, all in 8 1. of water. The animals were fed once in 2 weeks over a period of 3 months by dropping 100 g. of minced beef on each side of the wire netting. The movements of the toads while feeding were sufficient to keep the water well stirred, and the mesh of the partition was large enough to allow small pieces of meat to pass readily through from one side to the other. The water was changed once in 2 weeks, 24 hr. after feeding, and the food intake figures found as before. At the end of the experiment all the toads received the same dose of an extract of the anterior lobe of the pituitary. The number of toads which ovulated was recorded for each batch. The results are summarized in Table IIIB.

Table IIIB
graphic
graphic

They show that a major cause of the overcrowding effect is some form of interaction between the animals themselves apart from pollution. This may be due to fatigue as a result of competitive effort before they have eaten much food, or to some other sensori-motor phenomenon. None the less the antecedent experiment demonstrates that the maximum density level compatible with ovarian growth is higher if the water is kept running than if it is only changed after feeding.

Ovulation of Xenopus laevis in response to injection of extracts of anterior lobe tissue was first demonstrated by Hogben (1930), who had previously observed retrogression of the ovaries after hypophysectomy or removal of the pars anterior alone. In his preliminary communication and in a later publication (Hogben et al. 1931) little reference was made to the specificity of this reaction. In a private communication Hogben said that this did not seem necessary at the time, partly because various specificity tests were being carried out with regard to work on the chromatic function and partly because he attached more importance to operative removal of the gland. More recently the issue has acquired a new interest. Shapiro & Zwarenstein (1937) found that progesterone and testosterone induced ovulation. This was confirmed in the present investigation when the solid material was dissolved in alcohol and then diluted with water prior to injection. Neither “proluton” nor “perandren” were effective. This is possibly because they were not readily absorbed. Subsequent experiments have been undertaken to test the specificity of the response to injection of aqueous and alcoholic extracts of various bovine tissues. Doses equivalent to 5 g. of tissue were administered to toads taken from a batch which consistently responded to an aqueous extract of 0·2 g. of anterior lobe pituitary tissue. No result was obtained for either water or alcohol extracts of ovary, testis, adrenal, kidney, liver, pancreas, thyroid, muscle, brain, corpus luteum.

The Xenopus test was found to give positive results with menopausal urines. Urine of mares and ewes during the first 3 weeks of pregnancy elicited no response in doses equivalent to 50 c.c. urine.

For the benefit of those who wish to use Xenopus as a test animal the following is a brief summary of the technique suggested by the results of this and other investigations :

  1. Animals which have previously ovulated after injection should be used.

  2. They should be fed on liver once a week and kept at about 20° C. in running water or in a large volume of water (one 50 g. toad per litre).

  3. They should not be used for test unless at least 2 weeks have elapsed since their last ovulation.

  4. To test for suspected pregnancies of over a month’s duration the urine may be neutralized and shaken with a little ether to remove toxicity. 2 c.c. of the treated urine is then injected into the dorsal lymph sac.

  5. For suspected pregnancies of less than a month’s duration the urine must first be concentrated. The specimen is made slightly acid, made up to 80% alcohol and left to stand overnight at o° C. The active principle is precipitated by this treatment, and the fluid is then filtered through a Büchner funnel. The precipitate is washed on the paper with a few c.c. of ether. It is allowed to dry and then taken up with sufficient water so as to concentrate the original sample ten times. For very early cases the urine may be concentrated fifty times by this method. 2 c.c. of this solution is then injected into the dorsal lymph sac. Positive results have been obtained with cases only 5 days after the first missed period.

  6. Six animals should be used for each test and should be kept separately in glass jars. Since clawed toads often eat their own eggs the bottom of each jar should be covered with a mesh made by folding a piece of wire netting. The eggs drop through the mesh and the toad is unable to eat them.

  7. Xenopus responds to pregnancy urine in 12-18 hr. at 22° C. and 18-24 hr-at 18 ° C. If the test is negative after 24 hr. the animals may receive a further injection of the same sample (kept meanwhile in the refrigerator). They should then be left for a further 24 hr.

This investigation was started at the suggestion of Prof. Lancelot Hogben to whom the author’s best thanks are due for encouragement and advice. The progesterone used in this study was kindly supplied free by Schering, Ltd., and the testosterone propionate by Ciba, Ltd. Other expenses were defrayed from a grant made by the Medical Research Council.

Alexander
,
S. S.
&
Bellerby
,
C. W.
(
1938
).
J. exp. Biol
.
15
,
74
.
Bellerby
,
C. W.
(
1933
).
Biochem. J
.
27
,
615
.
Bellerby
,
C. W.
(
1934
).
Nature, Lond.,
133
,
493
.
Bellerby
,
C. W.
&
Hogben
,
L. T.
(
1938
).
J. exp. Biol
.
15
,
91
.
Crew
,
F. A. E.
(
1938
).
Private communication
.
Hogben
,
L. T.
(
1930
).
Proc. Roy. Soc. S. Afr
.
5
,
19
.
Hogben
,
L. T.
,
Charles
,
E.
&
Slome
,
D.
(
1931
).
J. exp. Biol
.
8
,
345
.
Orent
,
E. R.
&
McCollum
,
E. V.
(
1931
).
J. biol. Chem
.
92
,
651
.
Robson
,
J. M.
(
1934
).
Recent Advancer in Sex and Reproductive Physiology.
J. and A. Churchill
.
Shapiro
,
H. A.
&
Zwarenstein
,
H.
(
1933
).
Proc. Roy. Soc. S. Afr
.
Sept
. 20.
Shapiro
,
H. A.
&
Zwarenstein
,
H.
(
1934
).
Nature, Lond.,
133
,
762
.
Shapiro
,
H. A.
&
Zwarenstein
,
H.
(
1935
).
S. Afr. med. J. March
.
Shapiro
,
H. A.
&
Zwarenstein
,
H.
(
1937
).
J. Physiol
.
89
,
38P
.