Japanese scientists last year reported the results of an extraordinary experiment in animal nutrition. The mice they worked with could well have been the best fed in the history of research — not in terms of quantity of food, but in its quality.

On a typical day, one group of mice got to eat mixed rice with dried whitebait and green seaweed flakes for breakfast, together with cooked beans and miso soup containing the root vegetable taro and Japanese mustard spinach. Another group got bacon and eggs, toast and fluffy boiled potatoes. Lunch for one group could be simmered pumpkin and ground chicken, with a portion of cucumber and wakame seaweed with vinegar dressing. A different group of mice got a hamburger and salad.

Dinner was selected by the scientists from dishes including prawns with chilli sauce, Sichuan-style bean curd, fried Japanese horse mackerel, white radish and shimeji mushroom soup, sake-steamed clams and steamed Japanese seerfish. The mice ate from that kind of menu every week of their lives. There was no pudding, but probably no complaints.

The reason for all this gourmet cuisine was to recreate the typical Japanese diet from decades past, and to examine its impact on health. The long life expectancy of people in Japan has been attributed to the benefits of Nihon shoku or traditional diet, involving fermented foods that seem to boost the protective effect of harmless microbes on and in the body. As the food available in Japan has become increasingly Westernized, the effects on health are being questioned.

Hence the mouse study. Each group was given dishes from recreated daily menus for a typical Japanese household in 1960, 1975, 1990 and 2005. The food was ground up and fed to the animals along with their regular meals of ‘standard laboratory chow’. As the scientists suspected, the animals that were fed the older, more traditional diets lived and prospered for longest (K. Yamamoto et al. Nutrition 32, 122–128; 2016).

There are two things to note. The first is the large contribution that the environment — in this case diet — can make to health. The second is that such experiments enable the impact of environment on health to be assessed in ways that are simply not possible for human trials.

Light, heat, food, company, exercise, distractions, stress — all are at the fingertips of scientists who set up mouse experiments. Subtle changes in any of these can lead to profound, and potentially useful, discoveries about how health is changed by external factors. Research has probed, for example, how giving mice tunnels, stairs and wheels to play with can alter how female mice interact with their young, which in turn alters the brain development of the offspring (T. Begenisic et al Neurobiol. Dis. 82, 409–419; 2015). There is some evidence that modern, sterile, individually ventilated cages — used to minimize the effects of environmental factors such as disease — are quieter and less smelly for the mice, which reduces stimulation of those sensory systems.

Given that we know that environment affects the outcome of experiments, it is surprising that we don’t know more about the environmental set-up of other studies — those that test the impact of a potential medical treatment, for example. As we report this week, many researchers who use mice do not even know the content of standard lab chow and how it may change from study to study. As scientists hunt for sources of irreproducibility in their research, variation in living conditions — and how to reduce it — deserves more attention.