Conclusions: The association between geographical area and cervical cancer persists after controlling for 10 other factors.

The introduced pathogen Cronartium ribicola, cause of white pine blister rust, has spread across much of western North America and established known infestations within all but one species of white pine endemic to western Canada and the United States. Blister rust damage to severely diseased trees reduces reproduction and survival. Severe losses in white pine populations have resulted in site conversions to other species and seriously impacted resource values for timber, wildlife, watershed, recreation, aesthetic and other ecosystem services. In addition to blister rust, other major forest health threats and challenges to sustaining or restoring white pine populations are infestations of other pathogens, insects, fire, management practices that favour other tree species, and climatic change. Recent, large-scale outbreaks of mountain pine beetle have raised concerns for the viability of some white pine populations. In the 1960s, forest disease management for western white pine and sugar pine shifted from Ribes eradication to planting seedlings selected for better survival and resistance to blister rust. Seed orchards for producing improved white pines have been established, but deployment of that improved stock is hampered by a lack of planting opportunities. The inheritance and mechanisms of resistance are best known for western white pine and sugar pine; but new work is extending an understanding of genetics to all the western species of white pine. Current management efforts are focused on locating and protecting individual trees resistant to blister rust and assessing their disease resistance and other adaptive traits. In response to the threats from blister rust, the strategic goal is to sustain or restore viable white pine populations in western forest ecosystems. The four action components of the strategy are: (1) conserve genetic resistance to C. ribicola; (2) reduce the risk of adverse impact in stands currently uninfested; (3) restore and maintain white pines where blister rust is causing impacts and (4) assess and monitor the health and management of white pines. Successful implementation requires long-term support for coordinated efforts of management and research agencies, forest industry and an informed public.

Objective-To describe the characteristics of general practice patients who fail to respond to an invitation to attend for a health check, in relation to demographic variables, risk factor status, health status, and attitudes to behaviour modification. Design-Postal questionnaire before invitation to attend a health check and subsequent record of attendance. Setting-Five urban general practices in Bedfordshire, UK. Subjects-A total of 2678 patients aged 35-64 years were invited for a health check in 1989-90. Results-The number ofpatients who did not attend was low overall but was higher among men than women (21 v 15%, p<0 001), and in unmarried than married patients (24 v 16%, p<0-001 British general practitioners are now offered financial incentives to provide routine health checks for their patients. Many practices have established clinics, administered by nurses where a number of screening checks are carried out and advice is offered on health related behaviour, such as smoking, diet, exercise, and alcohol consumption.' The effectiveness of these checks in reducing risk, however, has not yet been shown. Whether this approach will reach those most in need of preventive advice has been questioned. The study of Waller et al indicated that attendance at these health check clinics administered by nurses was low and inversely related to the patients' cardio-The OXCHECK study is a randomised controlled trial of the effectiveness of nurseadministered health checks in helping patients to reduce their risk of heart disease, cancer, and stroke. Demographic, socioeconomic, and a variety of risk factor data were obtained by questionnaire from the study population before randomisation. Respondents were then randomised to be seen in one of four years of the trial, which is due to end in 1993. This, coincidentally, has provided an excellent opportunity to examine further the characteristics of those who do and do not respond to an invitation to attend for a health check. MethodsAll male and female patients aged between 35 and 64 years in five general practices in Luton and Dunstable were sent a health and lifestyle questionnaire before randomisation into the OXCHECK trial. The questionnaire included questions on health, risk factors, anxieties about health, and attitudes to making health promoting changes in lifestyle. Completed questionnaires were received from 11 090 of the original list of 17 965 patients taken from the family practitioner committee registers. After those patients who were known to have moved or died (n=2327) had been excluded, the reponse rate was 72-5%. We had, however, previously estimated that because of inaccuracies in the sampling frame, the true response rate was 80-3%.Questionnaire respondents were randomised to be invited for a health check in one of four years (mid 1989 to mid 1993

In the Sierra Azalea occidentalis, another charming shrub, grows beside cool streams hereabouts and much higher in the Yosemite region. We found it this evening in bloom a few miles above Greeley's Mill, where we are camped for the night. It is closely related to the rhododendrons, is very showy and fragrant, and everybody must like it not only for itself but for the shady alders and willows, ferny meadows, and living water associated with it.

Objective: To determine whether responses to simple dietary questions are associated with specific causes of death. Design: Self-reported frequency intakes of various classes of foods and data on confounding factors were collected at the baseline survey. Death notifications up to 31 December 1997 were ascertained from the Office for National Statistics. Relative risk (RR) of death and 95% confidence intervals (CI) associated with baseline dietary factors were calculated by Cox regression. Setting: Prospective follow-up study based on five UK general practices. Subjects: Data were used from 11 090 men and women aged 35-64 years (81% of the eligible patient population) who responded to a postal questionnaire in 1989. Results: After 9 years of follow-up, 598 deaths were recorded, 514 of these among the 10 522 subjects with no previous history of angina. All-cause mortality was positively associated with age, smoking and low social class, as expected. Among the dietary variables, all-cause mortality was significantly reduced in participants who reported relatively high consumption of vegetables, puddings, cakes, biscuits and sweets, fresh or frozen red meat (but not processed meat), among those who reported using polyunsaturated spreads and among moderate alcohol drinkers. These associations were broadly similar for deaths from ischaemic heart disease (IHD), cancer and all other causes combined, and were not greatly attenuated by adjusting for potential confounding factors including social class. Conclusions: Responses to simple questions about nutrition were associated with mortality. These findings must be interpreted with caution since residual confounding by dietary and lifestyle factors may underlie the associations. KeywordsDiet Nutrition Follow-up study Risk factorsThe complex relationship between diet and health is of enduring importance to consumers, health practitioners and policy makers. While the role of specific nutrients in the aetiology (and prevention) of common diseases is currently the focus of intense research 1 , there remains a gulf between defining these complex and interrelated biochemical pathways, and the practicalities of population-wide dietary modification. Few reports have been published relating disease incidence or mortality among industrialized populations to consumption of broad food groups (e.g. vegetables, red meat, milk, fruit), yet most nonnutritionists are likely to perceive their dietary intake in this way.To address the question of whether simple dietary assessments are associated with risk of death from specific causes, we analysed the mortality experience of a prospective study of more than 11 000 patients in a British primary care setting with 9 years' follow-up. Methods SubjectsThe study population consisted of all those who participated in the baseline data collection for the OXCHECK study, a randomized trial of nurse health checks in primary care. A detailed description of the protocol has been provided in an earlier report 2 . Briefly, in 1989 questionnaires wer...