4i6 



NATURE 



[February 28, 1895 



bacteria, amongst which »ere those of anthrax and cholera, 

 lived for several days, in the case of the former twenty-one to 

 twenty-four days, and in that of the latter thiity-five days and 

 even longer, typhoid bacilli were destroyed already in the course 

 of forty-eisht houri in sterilised sea-water. These results teiid 

 to contirm those obtained by Prof. Percy Frankland, who, in 

 his last report to the Royal Society Water Research Committee, 

 states that the addition of I and 3 per cent, of common salt to 

 ordinary Thames water into which typhoid bacilli were intro- 

 duced, acted very prejudicially on the latter, although it 

 stimulated the multiplication of many forms of water bacteria 

 picicnt in a most astonishing manner. Similar results were 

 obtained with salted but sterilised Thames water infected with 

 typhoid bacilli, and not only did the latter disappear very 

 rapidly, but the typhoid colonies which were obtained on plate- 

 cultures from this salt water exhibited in some cases a very 

 abnormal appearance, attributable to the degeneration of the 

 typhoid bacilli under these conditions, for on passing such 

 colonies through a further process of plate-culture, they returned 

 to their ordinary type. 



Giaxa also experimented with fish, introducing pathogenic 

 microbes /cr c.c. Unfortunately, he did not select the typhoid 

 bacillus for these experiments, but took instead cultures of 

 cholera and anthrax bacilli. He found that in both cases these 

 micro organisms were entirely destroyed in the course of a few 

 hours. Experiments were also made with oysters and some 

 varieties of mussel-fish. .\ small hole was bored in the shell, 

 and vigorous broth cultures of anthrax and cholera bacilli were 

 respectively introduced, after which the aperture was caicfuUy 

 closed with sealing-wax, and whilst some of the molluscs weie 

 replaced in sea water, others were kept in glass dishes without 

 any water. The latter lived for two days under these conditions. 

 Six hours after the injection had taken place, and again at the 

 end of twenty-four and forty-eight hours, the sealing-wax was 

 removed and a small quantity of the fluid within the shell was 

 taken out and plate-cultures made. In the majority of the ex- 

 periments these pathogenic microbes had completely disappeared 

 in six hours, whilst in only two instances were they detected in 

 small numbers at the end of twenty-four hours, and in no case 

 were they identified after forty-eight hours. These results were 

 irrespective of whether the moUu-c was kept in or out of water. 

 In these investigations, therefore, there would appear to be no 

 evidence that these pathogenic microbes (cholera and anthrax) 

 were capable of being transmitted by means of these jliell-fish. 

 .So far as I am aware, no one has followed up these interesting 

 experiments, and in view of the serious allegations which have 

 lately been made against oysters as transmitters of typhoid 

 fever, it is a subject which might well claim reinvestigation. 



February 13. G. C. Frankland. 



The Occurrence of very Cold Days. 



It is usual to estimate cold by minima of temperature. Hut 

 we may also consider it as expressed in maxima. A day may 

 fairly be called " very cold " in which the maximum is not over 

 freezing point. I propose to offer some account of the occur- 

 rence of such days at Greenwich in the fifty winter seasons 

 (November to March) 1S44-5 to 1S93-4. (For brevity, I will 

 designate each winter by the year in which it begins.) 



These very cold (l.iys are not very frequent. In some winters 

 there are none. The highest number is 27. The list is as 

 fallows : — 



The total, 251, gives an average of 5°02 for each season. 

 Ten of the winters had none of those days, and fifteen had 

 numbers over the average. 



The distribution in months w.is as follows : — 



Such days are rare in November ; and still more rare in 

 March. The six cases in the former month were in 1S49, 

 185S, 1861, 18S7, and two in 1S90 ; the earliest in the month 

 being that in 18S7 (16). The four cases in March were all in 

 one year, 1845. January h.is most of such days (about 17 per 

 cent, more than December). 



The largest number in any one month was sixteen in 

 December 1890 ; the next, twelve in January 1867 and January 

 iSSi. 



We might perhaps note that relative maxima of the seasonal 

 numbers occurred in 1854(15), in iSt')6 (12), in 1878(18), and 

 in 1890 (27) ; giving three intervals of twelve years. 



Measured by mean temper.iture of the three winter months, 

 these winters were all severe (considerably under average) except 

 I Sot) ; in which, though January (1S67) was very cold, December 

 and February were both mild (the latter uncommonly so) ; 

 thus the winter was above the average of mean temperature. 



Days witii Maxima not vier 32" at Greenwich. 



1840 '45 '50 'GS 'eO '65 '70 



00 '86 



NO. 



1327, VOL. 51 J 



Total 



251 



Thcie very cold days most often come in groups (two or mon 

 in succession). The larj;cst group was 10 in December '90 ; 

 the next, 9 in January '81 ; then 6 in December '44, December 

 '59, and January '93, kc. 



(If the 251 days, 100 had maxima from 31' to 32° ; 51 from 

 30° to 31 ; 42 from 29" to 30° ; 23 from 28° to 29" ; 14 from 27 

 to 28° ; and 21 under 27", ranging down to 19". The corrc 

 sponding minima range from 29 '6 down to 7''7, with grcalesi 

 density about 22" to 24° (45 cases). 



How are the maxima related to their respective minima? 

 There is, of course, no rigorous proporlionaliiy ; and we may 

 often find high maxima with very low minima {e.g. Dicemlier 

 29, '60, max. 3i'''8, min. lo°o), as also low maxima with rela- 

 tively high minima {e ff. January 15, '50, ni«. 26°-9. mi". 

 25°-9). On an average, however, high minima go with high 

 maxima, and low with low. Take the following : 



I selected the cases in which the maxima was under 27°; 

 there are 21. ' >f the corresponding minima only fmir were over 

 20". On Ihe other hand, taking the l.i>t 21 (of 46) c.iscs in 

 which the maximum was over 31 'S, one finds only lourminimB 

 iindir 20°. 1 he average minimum in the former case was l6°'4 ; 

 in the latter 24''4. I he average difference of maxima and 

 minima was in both cases about 8". 



