ABSTRACT
Two series of observations were made to determine the time and amount of loss of eggs in relation to the number shed per ovary and per mouse.
The correlations between sides within mice for eggs shed was −0·528, for corpora lutea counts −0·436, for implantations −0·300, and for live embryos −0·111.
A positive linear trend of loss of fertilized eggs with increasing numbers of eggs per uterine horn has been shown to occur before implantation.
Possible causative mechanisms for the loss are discussed in relation to observations on embryonic mortality previously reported by other workers.
We wish to express our gratitude to Prof. C. H. Waddington for laboratory facilities, to Dr D. S. Falconer for much helpful criticism and advice and to Dr B. Woolf for statistical advice. J. C. Bowman gratefully acknowledges financial support from the Agricultural Research Council.
INTRODUCTION
An inverse relationship between the numbers of foetuses in the two horns of the uterus in the house mouse during the last week of gestation has been reported several times (Hollander & Strong, 1950; Runner, 1951). We recently found a significant negative correlation between the two ovaries in respect of the number of corpora lutea present at 18 days of gestation. In respect of implantation sites and live embryos at 18 days, however, the correlations were much less strongly negative and were non-significant. This reduction in the strength of the correlation points to a differential loss of eggs or of embryos, the horn receiving the greater number of eggs suffering a proportionately greater loss. We have accordingly re-investigated these correlations and looked for direct evidence of a differential loss. We have found an increasing proportional loss of eggs related to the number shed within a horn, and have shown that it occurs after fertilization but before implantation. The results are reported below.
RESULTS
Two series of observations were made. In the first, pregnant female mice of heterogeneous origin were dissected at 18 days of gestation. The numbers of corpora lutea were counted as a measure of the number of eggs shed from each ovary. The numbers of implantation sites and of live embryos in each uterine horn were also counted. The correlation between sides within mice for corpora lutea was −0·436 (P <0·001). In agreement with our previous findings, the numbers of implantation sites and live embryos showed lower negative correlations between sides than the number of eggs shed. The correlation between sides for implantation sites was −0·300 (P<0·02) and for live embryos −0·111 (N.S.).
Further analysis revealed that the loss of eggs is affected by the number of eggs shed into the horn. Horns with a larger number of eggs suffer a proportionately greater loss. Thus, mice with an extreme distribution of eggs between sides will contribute largely to the negative covariance in corpora lutea counts, but as the two horns are affected by differential loss the negative correlation is greatly reduced by the time of implantation.
The distribution of loss of eggs and implanted embryos within uterine horns and within mice is shown in Table 1. The χ2 analyses for linear trend of proportion lost at two stages of gestation are given in Table 2. Only the linear trends of loss within horns up to implantation and for total loss up to 18 days of gestation are significant. The loss within horns between implantation and 18 days of gestation is very irregular and does not seem to be related either to the number of eggs shed into the horn or to the number of eggs implanted. We conclude therefore that before implantation there is positive linear trend of loss of eggs with increasing numbers shed per horn. This conclusion leads us to expect to find a similar increasing proportional loss when the data are analysed on a within-mouse basis. Our data does suggest such a trend, though statistical analysis reveals this to be non-significant.
The differential loss between horns up to implantation may be due either to a lack of fertilization or to a failure to implant, and a second series of observations was made to solve this problem. Females of heterogeneous origin, similar to those used in the first series of observations, were put singly with a male between 5.0 and 5.30 p.m. and examined the following day between 9.0 and 10.0 a.m. for vaginal plugs. Those which had mated were killed between 7.0 and 10.0 p.m. and the eggs in each Fallopian tube were extracted. From the findings of Snell, Fekete, Hummel & Law (1940), Snell, Hummel & Abelmann (1944) and Braden & Austin (1954) we considered that the majority of fertilized eggs would at that time be in the pronucleate stage. The eggs were examined by phase-contrast microscope according to the method described by Austin & Smiles (1948). The numbers of fertilized and non-fertilized eggs in each tube were counted. Judgement as to whether eggs were fertilized or not was based on the description by Austin (1951) of the formation of the pronuclei in the rat egg.
The correlation between ovaries in the numbers of eggs shed in a mouse was found to be −0·528 (P 0·001), which is in good agreement with the similar correlation for corpora lutea counts. The number of eggs not fertilized per tube in relation to the total eggs shed per ovary are shown in Table 3. In marked contrast to the implantation data, no regular trend of loss is apparent in this case. The data has also been analysed on a within-mouse basis and as expected no trend is shown. These results indicate that fertilization rate is not related to the number of eggs shed per ovary or per mouse and that it does not normally limit litter size in the mouse.
DISCUSSION
Our observations indicate that the loss of implanted eggs up to 18 days of gestation does not vary with the number of implantations in a horn, but that as the number of eggs shed into a uterine horn increases the probability of each individual egg implanting decreases. The fertilization rate is not related to the number of eggs in the horn, and therefore the factor or factors causing the variation in implantation rate must be operating on fertilized eggs or on pre-implantation embryos.
Our results and conclusions do not entirely agree with those of other workers. Danforth & de Aberle (1928) and McLaren & Michie (1956) found no correlation between the two horns for the number of implantations. As mentioned earlier other authors have reported significant negative correlations, and therefore we can only attribute the inconsistency of the results reported to heterogeneity between mice used at different laboratories.
The differential loss of eggs in our data could be explained if trans-uterine migration of eggs had occurred in many of our mice. Such migration is known to occur in rodents (Runner, 1951; Boyd & Hamilton, 1952; Young, 1953; McLaren & Michie, 1954), but these reports suggest that its frequency is very low. For this reason we have dismissed migration as an explanation of our results.
Previous work and ideas as to the causes of pre-implantational loss have been reviewed by Hammond (1952), but it is impossible to decide from our present experimental evidence which if any of the causes are applicable to our findings. Some useful conclusions may be made, however, by comparing our results with those of McLaren & Michie (1956).
From the results of an experiment, in which they transferred varying numbers of -day-old blastocysts from donor mice to normally mated recipient mice days pregnant, they concluded that ‘although we have found no limit to the number of eggs which can implant in a single uterine horn, we are beginning to approach a limit to the number of implantations which a single horn can keep alive’. In the data reported here the post-implantational loss was irregular and not proportionately related to the number of implants in the horn. The reason for this apparent difference is fairly easily found. McLaren & Michie (1956) consider that in their material the limit to the number of implantations which remain alive to 16 days of gestation may possibly be due to ‘insufficiency of corpora lutea to supply the progesterone requirements of the excessive number of implantations’. In all animals included in our data, presented here, there were at least as many corpora lutea in the ovary as implantations in the horn to which it corresponded, and consequently the progesterone supply is much less likely to have been insufficient. It is possible, however, that the total number of implants per mouse surviving to birth is limited by the level of some substance circulating in the maternal blood supply—a hypothesis favoured by Runner (1951) and by Hammond (1952).
McLaren & Michie (1956) found in their experiment that the number of successful implantations, from donor and recipient sources combined, rises linearly by increments of 0-2 for each additional egg injected. Again, this conclusion seems to conflict with our results of an increasing proportional loss of eggs up to implantation. However, the discrepancy may not be so serious as it first appears. If indeed there is no increase in the fractional loss in McLaren & Michie’s data, then it seems highly probable that the increased proportionate mortality in our material occurred very soon after fertilization, in fact between fertilization itself and the stage at which the blastocysts were removed by McLaren & Michie for transplantation.
If this suggestion is correct we have a more accurate estimate of the time interval during which the differential loss of eggs takes place, and this knowledge might prove useful in further elucidation of causes of pre-implantational loss.
ACKNOWLEDGMENTS
We wish to express our gratitude to Prof. C. H. Waddington for laboratory facilities, to Dr D. S. Falconer for much helpful criticism and advice and to Dr B. Woolf for statistical advice. J. C. Bowman gratefully acknowledges financial support from the Agricultural Research Council.