The relation between maternal genital colonization by mycoplasmas and fetal growth was examined in a study of 195 women. Swabs were taken from the endocervix on three occasions during pregnancy and once post partum. Ureaplasma urealyticum organisms (ureaplasmas) were recovered from 42.7 per cent of Caucasian women and from 34.6 per cent of Asian women at their first antenatal visit. These isolation rates remained similar throughout pregnancy, although there was a decrease in isolation after delivery. Mycoplasma hominis was recovered from 6 5 per cent of Caucasians and from 1 1 5 per cent of Asians at their first antenatal visit and these rates remained fairly constant during pregnancy and after delivery. Caucasian women colonized by ureaplasmas had a longer mean length of gestation (p < 0.025) than non-colonized women. Furthermore, the colonized women gave birth to infants who had a statistically significant greater mean birth weight and a greater mean birth weight-for-dates than those of the non-colonized Caucasians. There was no correlation between gestational length, birth weight, or birth weightfor-dates and genital colonization of Asian mothers by ureaplasmas or M . hominis. It is clear that ureaplasmas are not associated with low birth weight in our population.

Summary: A longitudinal study of the cervical and vaginal microflora of 131 pregnant women showed a similar range of organisms at first visit (up to 16 weeks), at 28 weeks and at 36 weeks of pregnancy. Twenty different groups or genera of microorganisms were recovered, predominantly lactobacilli. There was a fall in the isolation rate of organisms in the mothers 6 to 8 weeks after delivery. Lactobacilli and yeasts including Candida albicans were recovered less .frequently whereas the incidence of Escherichia coli and group B streptococci increased.Infants born to these mothers were swabbed within 24 hours of delivery and yielded organisms from the umbilicus, ear and mouth in 24%, 33% and 38% of cases respectively. Alpha haemolytic streptococci and Staphylococcus aureus were the-predominant organisms. Sixteen different groups or genera were isolated.The isolation rates in non-pregnant women attending a 'well-women' clinic were similar to those in the postnatal mothers; few women of 60 years or more were colonized.

SUMMARY Swabs were taken from the posterior fornix, perineum, and anorectum of 135 patients on three occasions during their pregnancy. Multiple isolates of /3-haemolytic streptococci of group B were obtained from 24 women, in 21 of whom the strains were examined by a highly discriminative serotyping and phage typing method. In 18 of these patients their own isolates were indistinguishable but different from those of other women with multiple isolates. Women yielding group B streptococci from the posterior fornix usually carried an indistinguishable strain in the anorectum.Haemolytic streptococci of Lancefield's group B are an important and possibly increasing cause of neonatal morbidity and death.1-3 Neonatal colonisation usually arises from the mother's genital tract.4 5 In this study we have used a highly discriminative serological and bacteriophage typing methods to examine the similarity or otherwise of strains of group B streptococci obtained from the vagina, perineum, and anorectum of pregnant women in an attempt to gather evidence for the source of this organism in the female genital tract. Patients and methodsThe 135 patients studied were randomly selected from the women attending the antenatal clinics of Northwick Park Hospital. Swabs were obtained from their posterior fornix, perineum, and anorectum at 16, 28, and 36 weeks of pregnancy. In some patients the skin of the anterior axillary fold and the mucosa of the urethra were also swabbed. The perineum and axilla were sampled by cottonwool swabs moistened in peptone water, the urethra by a narrow-gauge soft-wire swab, and the posterior fornix and anorectum by charcoal-impregnated swabs. Swabs were immediately broken off into 10 ml Todd Hewitt broth containing 8 ,ug/ml gentamicin and 15 ,pg/ml nalidixic acid and the broth incubated at 371C.7 8 Subcultures were made on to Columbia blood agar and incubated aerobically. Isolates resembling group B streptococci which were able to grow on 40% bile agar and failed to hydro-

As human genital mycoplasmas have been associated with various forms of reproductive failure, the present study was undertaken to investigate whether M. hominis and U. urealyticum organisms (ureaplasmas) are capable of crossing intact fetal membranes. Nearly 300 women in Denmark and England were investigated. Most of them were seen at about the fourth month of gestation and the remainder towards or at the time of birth, all with unruptured membranes. A swab was taken from the uterine cervix or vagina and M. hominis was isolated from 9% of the women and ureaplasmas from half of them. The presence of these mycoplasmas was not associated with an abnormal outcome of pregnancy. In contrast to the frequent presence of mycoplasmas in the lower genital tract, amniotic fluids obtained by transabdominal amniocentesis or at cesarean section did not contain M. hominis and ureaplasmas were isolated from only one of them. This was associated with the same ureaplasmas serotype being recovered from the cervix and also from the blood of both infant and mother, whose case differed from the others as labor had already started when the amniotic fluid was obtained. Thus, in our populations, we have no evidence that mycoplasmal invasion of the amniotic fluid occurs before the onset of labor. During labor, despite intact membranes, it seems that genital mycoplasmas may occasionally invade the fetal--placental unit, probably by the hematogenous route after strong uterine contractions, or otherwise directly after membrane rupture. Since both these events are followed usually by immediate delivery, there would seem to be insufficient time for the genital mycoplasmas to cause fetal damage.

An enzyme-linked immunosorbent assay (ELISA) was developed to measure serum antibodies to group B streptococci in 20 healthy pregnant women before delivery and in their newborn infants. The sera from 10 of these women who were colonised with group B streptococci and umbilical cord sera from their infants, had higher levels of type-specific IgG antibody than the 10 non-colonised controls and their neonates. All the babies remained well. The results demonstrate that infants from colonised mothers receive type-specific antibody. The possibility that this antibody may provide some degree of protection at birth against this potentially lethal organism warrants investigation.