The aim of this study was to characterise eD in the roe deer using non invasive transrectal ultrasonography. Females were monitored continuously during their breeding season to identify morphological changes of the genital tract and to assess embryonic and foetal development following implantation. Postmortem examinations ex situ and in situ were performed prioe to in vivo examinations in order to describe the sonomorphology of vagina, cervix, non gravid uterus and ovary. Ovarian and uterine changes visualised by Transrectal Adapter-Sonography (TAS) as monitored throughout the breeding season, eD and subsequnet development of the conceptus after implantation, are summarised as follows:
1. Oestrus-associated changes in the ovary, uterus, cervix and vagina were detected. On the ovary, follicles were identified retrospectively to have ovulated based on the topographical positioning of their corpora lutea shortly after oestrous. The development of lutein tissue, followed up sonographically, was usually completed 10 days p.o. Sonograms of uterus, cervix and vagina showed oestrus-associated oedemas. In comparison to histo-morphological changes of the endometrium as observed in cattle and small domestic ruminants the Tunica propria of the roe deer endometrium was oedemous and the uterine glands were highly coiled. In the early luteal phase a decreasing endometrial thickness was observed 8 - 10 days after mating, presumably a consequence of an increased progesterone stimulation. The end of luteal tissue proliferation coincided with the decrease of endometrial oedema and height. Sonographically observed ovarian and uterine changes during and shortly after breeding in roe deer including follicular growth, development of corpora lutea, oedema of the endometrium and changes of the endometrial epithelium resembled those observed in other ruminant species.
2. Development, prime and regression of single corpora lutea was followed sonographically during eD. Active corpora lutea in gravid and non gravid animals had, because of their plasma rich granulosa-lutein-cells, an anechogenic appearance similar to large antral follicles. Furthermore, the number of implanted embryos did not correlate with site or number of detected corpora lutea. In December, regression of corpora lutea was sonographically observed both in gravid and non gravid animals. This observation of corpus luteum regression independent of the animals´ reproductive status explained results of morphological studies, which showed no correlation between C.l. number and number of implanted embryos in roe deer. In none of the animals examined any development of accessory c.ll. during eD was observed. This supports the hypothesis, that in roe deer, progesterone is also synthesised in the placenta and, therefore, is responsible for the rise in plasma progesterone following implantation. Concluding from variations in number and size of antral follicles during eD it is postulated that follicles undergo an undulating growth pattern and seems to be responsible for fluctuating oestrogen level during this period. From the third month of eD onwards differences in gravid and non gravid animals in the sonomorphologic appearance of the endometrium were detected. The endometrium of gravid individuals was characterised by an increase in fluid associated grey levels. Long term ultrasound examinations revealed a continuous increase in endometrial fluid content over the last three month of eD, probably caused by a retention of secretory vesicles in cells of the uterine glands. During elongation of the trophoblast following eD the endometrium lost its dark echotexture and could hardly be distinguished from the adjacent, high echogenic embryonic membranes.
3. Since the human eye´s ability to distinguish different grey levels is very limited, quantitative greyscale analysis was performed to assess endometrial echotexture during eD. Quantification of endometrial echotexture made differences obvious, which to the examiner were otherwise barely visible in sonographic images. The grey-scale analysis of the endometrium quantified the increase in fluid associated grey levels in the endometrium of pregnant animals starting in the third month of eD. Mean medians of grey-level histograms in pregnant roe deer were significantly lower than in non pregnant animals. Early pregnancies diagnosed by computer-assisted tissue characterisation were later on all confirmed to be pregnant by detection of embryos in early January. The analysis of grey-level distribution in addition to transrectal ultrasound was adequate to describe endometrial changes in the last three month of eD. Since embryonic structures cannot be detected before the 6th month of pregnancy in roe deer and progesterone-/ oestrogen-levels rise long after implantation in January, we evaluated a potential procedure to monitor pregnancy at a time pregnant and non pregnant individuals can not be distinguished endocrinogically.
4. Starting in January, rapid embryonic and foetal development was followed by the growth of the embryonic yolk sac (1 - 2 mm), of embryonic structures and a well-developed foetus. Integrity of foetal structures and foetal membranes, detection of heart beat and movements were obvious life signs, which were clearly distinguishable from the sonographical appaereance of embyronic death. Out of 44 animals examined at the end of January in a free ranging population 75 %proved to be pregnant. The animals showed great variations in the conceptus size and stage of development. Comparing the mean crown-rump length and the biparietal diameter of animals from geographically different populations no significant difference in the stage of development between populations was observed. Differing stages of development at this time were explained by varying (consecutive) mating dates within social age groups. In later stages of pregnancy transcutaneous ultrasound permitted better survey sonogram and better assessment of foetal structures, due to growing size and number of placentomes.
5. Embryonic death was visualised at different stages of early pregnancy in roe deer. Based on endometrial changes as observed in pregnant animals during eD, an embryonic resorption before implantation was observed. The endometrium of one animal changed from November to December from being low echogenic to high echogenic, different from what was observed in pregnant animals in November. Quantitative greyscale analysis of the endometrium in this animal in November and December showed a significant change in echotexture. It was concluded, that a complete embryonic resorption of one or all conceptuses before implantation had occurred. Of 44 free ranging animals examined in late January, 33,4 % (n = 11) showed signs of embryonic mortality. Embryonic death was detected in 12,1 % of all single pregnancies and in 6,1 % of all twin pregnancies, concurrent with another intact conceptus. In 15,2 % no intact conceptus could be visualised. Embryonic death was characterised sonographically by the absence of foetal vitality, lack of integrity of embryonic structures and membranes. Although in one roe deer endometrial changes had been detected and quantified during eD and embryonic and foetal structures had been visualised over the following months, a foetal abortion was documented in April.
