SUMMARY. Twenty-eight pre-term babies of low birth weight were monitored for developing microflora in throat, stomach and faeces during the first 3 weeks of life. The flora at all levels of the gastrointestinal tract differed from that of healthy breast-fed and artificially fed full-term babies. Colonisation of throat and stomach was delayed beyond 4 days of life in 87% and 60% of babies respectively. Only 10% of babies had "normal" oral flora throughout the period of study. Flora of the stomach was sparse, and resembled faecal flora. Faecal flora was established more rapidly than throat or stomach flora, and 70% of babies were colonised during the first 4 days of life. Initially Bacteroides spp. were predominant (57% babies), but Escherichia coli and other aerobic gram-negative bacilli gradually increased in frequency. Colonisation by gram-positive bacteria was slow. Clostridium spp. were present in only 10% of babies during the first 4 days of life. Most strains were transient. Colonisation with C. butyricum (30%), C. perfringens (35%) and C. dzficife (25%) was maximum after the first 2 weeks of life. Lactic-acid-producing bacteria usually appeared late in the third week of life. Parenteral feeding immediately after birth was associated with delayed colonisation by a restricted number of species. Parenteral antibiotics (penicillin or gentamicin or both) restricted colonisation with normal oral flora, the lactic-acid-producing bacteria and penicillin-sensitive clostridia, but had little effect on E. coli even when the colonising strain was sensitive to the aminoglycoside in the regimen. Systemic spread of bacteria via the blood stream was not detected in any babies.The pattern of colonisation of the enteric tract in pre-term infants in the special-care nursery studied, differs from that of healthy full-term babies; this merits consideration when the results of bacteriological tests on this vulnerable group of infants are being interpreted.

Among 884 hospital patients with gastroenteritis, 36 (4.1%) were excreting Cryptosporidium oocysts in their stools; only 5 of the 36 patients were also excreting other enteropathogens, while none of 320 hospital patients without gastroenteritis were excreting Cryptosporidium oocysts. Children were more commonly infected with Cryptosporidium (4.8%) than were adults (1.6%). The prevalence of infection was higher (7%) during the summer period of February-May 1981 than in the remainder of the observation period to the beginning of June 1982 (1.9%). The most common clinical manifestation of gastroenteritis in Cryptosporidium-infected patients was diarrhea, lasting from 3 to over 14 days, accompanied by vomiting, anorexia, and abdominal pain. The results show that a small proportion of patients with gastroenteritis are infected with Cryptosporidium, and the importance of the infection needs to be examined.

Health professionals frequently write at the same level for lay readers as they write for peers. In relation to health research and ethical requirements to provide written explanation of studies, this can complicate the notion of informed consent. Plain language information statements need to be clearly understood by research subjects if the ethics process for research approval is to fulfil its objective. Many delays in gaining ethics approval for child-related research are caused by substandard plain language statements (PLS). We describe specific issues for information statements for research with children, young people and their parents/guardians, particularly in consideration of the literacy capabilities of the general population. We highlight the usefulness of everyday language when explaining research and science in writing to families, and present some guidelines for writing PLS that have emerged from the introduction of a plain language service by an Ethics in Human Research Committee.