The present review furnishes evidence that paratyphoid fever exhibits a good many differences from enteric fever in its epidemiological features.

The incubation period is decidedly variable and, in general, shorter than it is in enteric fever. Not only is the disease usually milder, but the case mortality figure in the present series is particularly low and one more comparable to the case mortality in Salmonella food poisoning than that of enteric fever.

It has to be recognized that acute gastro-intestinal symptoms may be present, very rarely indeed associated with all the cases but more often affecting a proportion of them, which so closely simulate acute food poisoning that that condition may be diagnosed. Usually, however, the same patients later exhibit the normal enteric picture.

The seasonal prevalence, at least as regards the onset of multiple outbreaks as distinct from isolated cases, shows a definite summer prevalence more comparable to the distribution of Salmonella food poisoning than to that of enteric fever.

The vehicles of infection are markedly different and, apart from outbreaks spread by case-to-case infection as in some institution outbreaks, almost invariably the vehicle is some form of food. Whereas milk and ice cream are common agencies in spreading both diseases water and shellfish are absent, or very rare, as vehicles in paratyphoid fever. For this disease cream takes pride of place.

I suggest that these conspicuous differences of vehicle are not associated so much with lack of opportunity to infect as with the need for a vehicle in which multiplication of the paratyphoid bacillus can occur previous to ingestion. These differences of vehicle are explainable on the assumption that the paratyphoid bacillus is less efficient at invading the human body than the typhoid bacillus, and therefore only vehicles which ensure its ingestion in considerable numbers are effective. The few water infections cited in the report bear out this contention, for in all of them water was only effective when massively infected and repeatedly consumed. This explanation is, of course, only a hypothesis, and it may be that other factors, such as preformed toxin, are also required for successful invasion.

The special implication of cream, it may be advanced, is partly due to this factor, and also that this vehicle is one which affords special opportunities for specific infection from an infective food handler.

The vehicle infectivity rate is definitely low, and in many outbreaks a surprisingly small proportion of persons at risk are attacked. The comparative infrequence of definite explosive outbreaks, even with a vehicle from which we should anticipate an outbreak to be explosive in character, is noticeable. Both these facts add support to the view that the paratyphoid bacillus does not readily infect man.

As regards the sources of infection of the vehicle I suggest that the importance of the chronic carrier has been overstressed, and that such carriers are less important in paratyphoid fever than in enteric fever. Our information is still inadequate, but the facts incline one to believe that the true chronic carrier is proportionately rarer than in enteric fever but is likely to be more associated as a causal agent with individual infections where there are opportunities for repeated infections.

On the other hand, my series of outbreaks emphasizes the fundamental importance of the transitory carrier who may be either a mild unrecognized case or a symptomless bacterially infected person. It is evident that in most outbreaks there is in addition to the recognized and so notified cases a reservoir of infected persons in one or both of these groups. Unrecognized by themselves and unknown to the Health Authority they constitute a pool of potential infection, the effectiveness of infection only limited by opportunity, and, to a minor extent, by the difficulty of the paratyphoid bacillus to establish itself as a cause of disease in the human body. No system of control can afford to neglect the potential menace of these two groups of persons.

Much more attention might profitably be given to a detailed study of the extent to which bacillary infection occurs without recognized symptoms and the extent to which these cases clear up within a reasonable period. A comprehensive investigation of isolated single cases of paratyphoid fever would, I am sure, yield valuable epidemiological results.

Influenced no doubt by the prevailing emphasis on the chronic carrier as the commonest source of infection my series shows a tendency to label any person who handled the infected vehicle and whose stools contain the paratyphoid bacillus as the source of infection of that vehicle without steps being taken to prove the contention. So frequently can the presence of the specific bacillus be equally well accounted for on the assumption that the person is a victim of infection in the actual outbreak that some attempt at proof is essential. Modern bacteriological methods are in a position to throw considerable light on the distinction between a potential cause of the outbreak and a victim of that outbreak and ought invariably to be employed.

I am impressed with the extent to which the undoubted fact that Bact. paratyphosum B is a true Salmonella influences the epidemiological picture, and I have suggested that a study of it from this angle furnishes us with a rational explanation of the main differences between its epidemiological behaviour compared with that of the typhoid bacillus.