Again, as in the case of one-spike fever many different viruses can be isolated from childrn showing roseolarreactions. There is, however, some seasonal variation in the appearance ofsuch rashes and it does at present seem probable that they are more often associated with intestinal rather than respiratory viruses. Once moreI would underline the value of family studies. In one household a child of 13 months developed a roseolar reaction, shown to be due to Coxsackie virus type A6. From her 2j-year-old sister the same type of virus was isolated but she only developeda one-spike fever. Having failed to resist infection or to achieve either a symptomlessimmune reaction, a fast one-spike immunizing fever or the slower roseolar reaction, a certain proportion of those exposed to any viral pathogen will develop so-called typical symptoms of illness. The proportion who do so will vary according to many unknown and some measurable factors, including age and sex, as shown in a series of secondary attack rates among home contacts of rubella (College of General Practitioners 1963). Not only is rubella remarkably uncommon among adult male contacts, which may be because they are out of the house so much, but also at all ages over 5 years rubella is significantly commoner in females than males and this cannot be solely due to lack of exposure.

Just as we saw that the febrile reaction is maximal in young children so too is the imnmunological response, for example, to a dose of live measles virus vaccine. This raises a number of problems, both clinical and administrative, in selecting the optimum age at which to administer measles vaccine (Watson 1967). Children in their