View report page

2
Cannot be said against sewage-grown land plants," for in the former the whole plant is liable to be
bathed in the sewage and "may even become encrusted with a deposit that careless washing will not
wholly remove." Again, he quotes the statement of Mr. Alfred Smee, who found that "watercresses
act as a scrubbing brush to the sewage and remove all the solid flocculi from the water, which adhere
to the stalks."
Some growers of watercress use artesian well water, and this not only on account of its purity,
but also because the warmth of the water enables cress to be grown and gathered in the winter months,
for brown cresses grow best in water having a temperature of 50° to 55° Fahr., and thus,
provided water of this temperature can be obtained all the year round, winter and spring cresses can
be produced. Sewage cress and impure river cress, on the other hand, grow best in water having a
temperature of 60° Fahr., and upwards, i.e., only in the summer months. "Summer cress" is, as a
rule, planted in April and early May, winter or spring cress in September, or, in sheltered positions,
even earlier. Sewage cress is not available for cutting during the early months of the year.
Enteric fever, is of course, the disease which is especially apt to be spread by sewage-contaminated
water in this country, and the suspicion that uncooked vegetables have been instrumental in promoting
infection has from time to time been entertained both here and on the Continent, more particularly
in France and Germany. Watercress has apparently in particular instances come into question in
this connection. The only observations bearing upon the matter, so far as the London population is
concerned, to which I am able to refer, are those made by Dr. Verdon, the medical officer of health
of Lambeth, in 1894, and by Dr. Warry, medical officer of health of Hackney, in 1903. Dr. Verdon
in the former year found that " a number of persons who were suffering from the disease (enteric
fever), in one of its severest types had, during the days coming immediately before the advent of the
earliest symptoms, eaten watercress on one or more than one occasion." Analysis of the water
supplying the beds from which some of the watercress in question had been supplied showed that
it had " all the chemical characteristics of liquid sewage." The evidence in this case, to quote
Dr. Verdon's own words, " begins with the incidence of typhoid, proceeds by indicating the consumption
of watercress at a period when infection must have taken place, continues by tracing the article
under suspicion to the shop of a local tradesman, and thence to the place of its growth, and ends by
demonstrating the existence of a connection of contact between the watercress consumed and sewage,
with its complement of living microbes." This evidence is no doubt suggestive, but in the absence
of material for precise statistical study is not, of course, conclusive.
The results of Dr. Warry's enquiry I am able to give in some detail. He noted the existence
of an unwonted prevalence of enteric fever in June, 1903, that is to say, before the time of year when
the usual autumnal increase would be expected; the cases notified were, moreover, for some weeks
"mostly localised within a circle of one-third mile radius." "These circumstances," he says, "led me
to think that some special agent was responsible for the unusual prevalence of this disease in the borough
which required special investigation." On further enquiry it transpired "that the outbreak took the
form of two waves or outbursts, the first extending from the second week in June to the week ending
the 25th July, and the second from the first week in August to the week ending the 29th August, both
of these outbursts showing a different localisation." These outbursts Dr. Warry distinguishes by the
letters A and B.
Outbreak A included 48 cases in 45 households; 29 of the patients, i.e., 64 4 per cent.,
had partaken of watercress, in most cases more than once during the three weeks prior to
the onset of illness.
Outbreak B included 62 cases in 57 households; 28 of the patients, i.e., 48.2 per cent.,
certainly partook of watercress during the three weeks prior to the onset of the disease.
In order to test the value of this evidence Dr. Warry caused enquiries to be made at 124 houses
"in those parts of the borough invaded by enteric fever," and ascertained " that the inmates of 34 only
were watercress eaters." "Thus," he says, "the watercress eaters may be taken at 27.5 per cent.,
and non-watercress eaters at 72.5 per cent, of the population; but the incidence of enteric fever on
watercress eaters compared with non-watercress eaters during this epidemic was as the figures 64: 36."
He further calculates "that watercress eaters suffered 4.6 times more than non-watercress eaters during
the A period, more than twice as much during the B period, and over three times as much during the
whole epidemic." Upon these statistical data and in the absence of evidence implicating milk, water,
shell-fish, ice-creams, and fried fish, concerning all which possible sources of infection careful enquiry
was made, Dr. Warry concluded that the watercress was at fault.
The apparent undue incidence of enteric fever on watercress eaters during the periods in question
is undoubtedly suggestive, but in considering how far the evidence available must be regarded as proof
that the watercress was the cause of the disease, the following considerations must be borne in mind.
In the two outbreaks 110 cases of this disease occurred in the areas in question. The number of cases
which would have occurred in the areas in the absence of an exceptional cause would be very few,
probably less than half-a-dozen. If this estimate be correct, this would still leave, after deducting
the cases occurring in watercress eaters as well as the half-a-dozen cases, an excess of some
fifty cases not explicable on the watercress hypothesis, unless, indeed, they could be regarded
as secondary cases, i.e., caused by personal infection from cases directly due to the consumption
of watercress. In this connection, it may be noted that so far as persons in invaded houses
are concerned, only in two or three instances were more than one person in the household known to
be attacked. The rarity of multiple attacks in houses, whether due to primary or secondary infection,
is noteworthy when it is recollected that if watercress is eaten in a household it is usually shared