SEPTEMBER 1931. Vol. LVI. No. 666, Carbon Monoxide Poisoning : Its Detection and the Determination of Percentage Saturation in Blood by Means of the Hartridge Reversion Spectroscope. BY ROBERT C. FREDERICK A.I.C. INTRODUCTION.-~O much has been heard in recent years of carbon monoxide poisoning that there is a tendency to regard casualties from this cause as being peculiar to the age. Though cases are now apparently more numerous the number has always been large and Lewinl goes so far as to state that carbon monoxide is now and has been since the first discovery of fire the most widespread poison connected with human life and activity. The vehicle in which the poison is conveyed has however certainly changed with the times. In the literature nearly a century ago there is discussion regarding the many fatalities occurring in France (a large proportion of which were suicidal) due to the use of braziers.French writers of fiction in their works frequently introduced 56 562 FREDERICK CARBON MONOXIDE POISONING situations where death took place by this means and one Zola himself met his death accidentally in this way. The use of fire fumes was a method of suicide employed by the Romans. At the present time in this country the most common cause of carbon monoxide poisoning is the public gas supply and history is repeating itself for the majority of these cases are suicidal. The extremely poisonous character of the gas is due to the carbon monoxide content being increased by admixture with water-gas and similar products. The Report of the Departmental Committee appointed to consider the increase in the number of deaths ascribed to poisoning by coal gas supplied for domestic purposes2 reveals that such as were suicidal increased from 193 in 1918 to 1191 in 1928; the figures for those considered officially as due to accident were 102 and 168 respectively.Apart from the question of gas escape poisoning may occur when the gas is burning if the flame is striking a cold conducting material; hence the menace of geysers and other types of hot water heaters when fitted without adequate flues. The advent of the petrol engine has brought in its train numerous cases of poisoning not all accidental due to the carbon monoxide in the exhaust gases. A certain number of cases of poisoning by carbon monoxide have occurred with men seeking warmth and falling asleep by the slow-burning heaps of town and ironworks refuse.Poisoning by carbon monoxide is an occupational hazard in many industries amongst which are water gas producer gas and coal gas manu-facture; blast furnace and coke oven operatim; lime and charcoal burning; and coal mining. The physiological response of the individual to various concentrations of carbon monoxide has been studied by Henderson Haggard Teague Prince and W~nderlich,~ and their findings have been summarised by the two first-named in the following table#* :-Concentration. Carbon monoxide. Per Cent. Allowable for an exposure of several hours . . 0.0 1 Can be inhaled for one hour without appreciable effect . . . . O-OPO-05 Causing a just appreciable effect after one hour's exposure .. . . 0*06-0*07 Causing unpleasant but not dangerous symptoms after one hour's Fatal in exposure of less than one hour . . 0.4 and above exposure . . . . 0.1 -0.12 Dangerous for exposure of one hour . . . . 0.15-0*20 The poisonous effect of carbon monoxide is exercised through the blood owing to the great affinity of the gas for haemoglobin (some 300 times that of oxygen) which results in the formation of carboxyhaemoglobin. The carboxy-haemoglobin is entirely incapable of carrying oxygen to the tissues and according to the amount of haemoglobin thus inactivated (or percentage saturation) so the individual exhibits illness of proportionate severity; if the percentage satura-tion becomes sufficiently high death ensues.The average physiological effects * Modified to show figures in terms of percentage instead of parts per million FREDERICK CARBON MONOXIDE POISONING 663 caused by varying percentage saturations of the blood with carbon monoxide have been succinctly set out by Henderson and Haggard in the following table6:-Haemoglobin in combination with carbon monoxide. Per Cent. 10 20 30 40-50 60-70 80 Over 80 Physiological effect. No appreciable effect except shortness of breath on vigorous muscular No appreciable effect in most cases except short wind even on Decided headache ; imtation ; ready fatigue ; disturbance of judgment. Headache confusion collapse and fainting on exertion. Unconsciousness ; respiratory failure and death if exposure is long continued.Rapidly fatal. Immediately fatal. exertion. moderate exertion; slight headache in some cases. Some of the differences between normal blood and blood containing carbon monoxide are readily evident such as the colour (which is invariably but rather unhappily described as cherry-red) whilst others are revealed only by the spectro-scope. The colour of the blood at least in the cadaver is not specific as is so generally supposed and it is essential in the absence of other positive evidence that spectroscopic examination should be made before arriving at a conclusion. Normal blood shows two absorption bands between the D and E lines and blood containing carbon monoxide will appear to exhibit identical features unless an instrument of precision is employed when especially if the percentage saturation of the blood is high and the spectrum of a normal blood is viewed simultaneously, careful observation will reveal that the bands of carboxyhaemoglobin are very slightly nearer the violet end of the spectrum.The spectroscopic difference is too slight to be diagnostic unless the Hartridge Reversion Spectroscope is employed. A chemical method of differentiation is available by observing the effect of the addition of a reducing agent such as ammonium sulphide when in theory, with a normal blood the two bands of oxyhaemoglobin merge into the one broad band of haemoglobin and in blood containing carbon monoxide the bands of carboxyhaemoglobin remain unchanged. This chemical method of differentiating the bands has been transcribed from one text-book to another and the accounts nearly all give the same impression of a process without difficulties or limitations.I n practice in my hands this test has frequently yielded results which were open t o doubt except with blood almost saturated with carbon monoxide even when carried out with simultaneous treatment and observation of a normal blood control. This is only to be expected for in the presence of a reducing agent the stable absorption bands of carboxyhaemoglobin are masked to a less or greater extent by the broad band of haemoglobin from reduction of oxyhaemoglobin according to the percentage saturation of the blood. With the Hartridge Reversion Spectroscope it is possible in a few minutes not only to detect carbon monoxide in blood with certainty even in small amount but also to determine the percentage saturation and therefore to express an opinio 664 FREDERICK CARBON MONOXIDE POISONING whether this was the cause of illness or death.In life the gas is rapidly eliminated as soon as the person is removed from the poisonous atmosphere but specimens of blood taken from an individual will retain the carbon monoxide for weeks if these are stored under suitable conditions. THE HARTRIDGE REVERSION SPECTROSCOPE Fig. 1. THE HARTRIDGE REVERSION SPECTROSCOPE.-The spectroscope is shown in Fig. 1; it is quite a small instrument measuring at the longest diagonal of the base only about 14 cm. A source of light must be provided at the collimating lens (seen prcjecting behind the micrometer screw).The lighting Enit (which may be obtained with the instrument) is conveniently a 60 C.P. electric lamp housed in one box arrangement superimposed on another which contains an adjustable mirror and has one side (facing the spectroscope) made of ground glass. I had this re-constructed and the complete apparatus with the spectroscope adjustable in position mounted on a miniature table 35 cm. high. Interposed between the source of light and the spectroscope is a holder for the cell containing the blood under examination. The cells used are 40 mm. square at the face and are of the type made by cutting a U-shaped piece from a solid piece of glass. Those made for me are 5 mm. thick; the depression is a true semi-circle (17 mm. in diameter at the bottom) and has an extreme height of 30 mm.The diameter is just sufficient to cover completely the collimating lens and enables an examination to be made if necessary with the absolute minimum quantity of sample-a single drop of blood. The principle of the instrument is the utilisation of the fact already referred to that in blood containing carbon monoxide the absorption bands are situate FREDERICK CARBON MONOXIDE POISONING 565 slightly nearer the violet end of the spectrum and in addition that the extent of this difference in wave-length is related to the percentage saturation. To enable this difference to be measured more accurately the effect is doubled by the simple and ingenious expedient of employing two spectra in reverse directions. With the instrument at zero the field of the spectroscope in observation of a normal blood is shown (Fig.2) in a and with blood containing carbon monoxide in b. By turning the micrometer screw the bands in either case may be brought into the basal position shown in c and a record obtained directly in Angstrom Units of the difference in wave-length between a normal blood and a blood sample containing carbon monoxide. should be used in a dark-room. The mirror of the lighting unit and the spectro-scope having been adjusted to give the essential uniform illumination of the spectra the reading for a normal blood is first obtained. The normal blood should be human and from an individual who is not a heavy smoker; Hartridge6 has found that blood from such subjects may be saturated with carbon monoxide to the extent of 6 per cent.Animal blood is to be avoided, as the intestinal gases of herbivora are stated to contain carbon monoxide; Hartridge’ has found carbon monoxide in sheep’s blood. The blood is diluted with distilled water* to such a degree that trial observations show the bands to be visible distinctly and the distance between them to be about the same as the width of the alpha band. For this purpose the small quantity of diluted blood required is pipetted (using a drawn-out glass tube with rubber teat attached) into or from a cell which is of course interposed between the instrument and the source of light. It will be necessary probably to focus; this is done by adjustment of the milled screw on the collimating lens. Once the position of this has been found it must not be moved again before examination of the suspected sample.Observation is made at the eyepiece seen on the extreme left of Fig. 1; it is fitted with a shield not shown. While these trial observations are being made the micro-meter screw (seen on the extreme right of the illustration) should be at zero. The knife edge in the centre foreground is a coarse adjustment but this is not used; it is placed exactly at 3 and is thereafter ignored except to see that it remains in that position. These preliminaries having been completed the reading for normal blood is now determined by turning the micrometer screw until the two alpha bands are in line (Fig. 2 C) and then when this position has been attained noting the figure on the screw.It is necessary to take the average of ten readings; a certain degree of variation will be found between the separate figures but the average of a series should be within one or two digits of another series of observations on the same blood. The most concordant results are obtained by making the adjustment quickly and without hesitation. The screw is graduated into 100 divisions each * The effect of adding distilled water is to burst the envelopes of the red blood corpuscles (hence the turbidity produced) and to cause the contained haemoglobin to pass into solution; this phenomenon is known as ‘ I laking.” This difference Hartridge terms the “ span.’’ THE DETERMINATION OF THE SPAN OF A BLOOD SAMPLE.-The instrumen 566 FREDERICK CARBON MONOXIDE POISONING representing one hgstrom Unit; these are marked at each tenth with the single numerals 0 to 9.If the adjustment of the bands entails the screw being turned away from the operator the graduations have to be read backwards e.g. a complete turn and the screw at 81 would be a reading of lOO+(lOO-81)=119. The suspected sample is now examined in precisely the same manner; it is a useful precaution thereafter to confirm the normal blood reading with a repeat series of observations. It is to be noted that the glass containers must not be sealed by fusion as carbon monoxide may be absorbed from the heating agent. The samples should not be exposed unduly to daylight as this may cause a loss of carbon monoxide. If the sample gives a reading appreciably greater than that of the normal blood i.e.the span is a definite figure the presence of carbon monoxide to the extent of a certain percentage saturation is revealed. Blood saturated with the gas gives a span of about 65 Angstrom Units. To determine the percentage satura-tion indicated by a span it is necessary to construct a calibration curve for the particular instrument employed as the figures are not in direct proportion FREDERICK CARBON MONOXIDE POISONING 567 PREPARATION OF CALIBRATION CURVE.-For the preparation of the curve it is necessary to establish four points the span of blood saturated with carbon monoxide to the extent of (i) 25 (ii) 50 (iii) 75 and (iv) 100 per cent. with carbon monoxide. Stock quantities of diluted normal blood (N) and blood saturated with carbon monoxide (S) must first be prepared.For N a small quantity of blood is placed in a stoppered bottle and distilled water is added to this until observation of a portion shows the degree of dilution to be such as has been specified in the foregoing. For S coal gas is passed through a similar quantity of blood in another stoppered bottle which is shaken from time to time; this blood is then diluted as requisite and afterwards again treated with coal gas to make certain of complete saturation. The reading for normal blood, which has to be deducted from each reading of blood containing carbon monoxide, is obtained by filling one cell with dilution N and the other with distilled water and taking the average of a series as already described. (i) Span for 25 Per Cent. Saturation.-In a test tube are placed 7.5 C.C.of dilution N and 2.5 C.C. of distilled water; in another 2-5 C.C. of dilution S and 7.5 C.C. of distilled water. Both dilutions are mixed in their separate tubes and a quantity from one is pipetted into one cell and from the other into another cell. With the two cells face to face as before the average of a series of readings is obtained, and this less that for the normal blood is the figure required. The other figures are determined by the same procedure except that the preparation of the test tube dilutions is varied as noted below. (ii) Span for 50 Per Cent. Satwation.-Five C.C. of dilution N and 5 C.C. of distilled water in one; 5 C.C. of dilution S and 5 C.C. of distilled water in the other. Two 5 mm. cells are employed face to face.(iii) Span for 75 Per Cent. Saturation.-The dilution N (2.5 c.c.) and 7.5 C.C. of distilled water in one; 7.5 C.C. of dilution S and 2-5 C.C. distilled water in the other. (iv) Span for 100 Per Cent. Saturatzout.-Distilled water aione in one; ciiiution S alone in the other. On plotting out the span figures a smooth curve should be obtained and this is extended to the zero point. The calibration curve once constructed is available permanently for immediate conversion of the span of a blood sample into terms of percent age saturation. NOTE ON THE PRESENCE OF NITRIC OXIDE HAEMoGLOBIN.-The possibility of NO-haemoglobin formed after death being mistaken for CO-haemoglobin in the examination of blood has been discussed by Banham Haldane and Savage,s following a case where the individual had not been exposed to carbon monoxide, yet the blood of whom post-mortem responded to certain of the usual tests for CO-haemoglobin ; the spectroscopic examination it is important to note was not carried out though they considered that had this been done it would have con-firmed the apparent presence of carbon monoxide.The opinion was expresse 568 FREDERICK CARBON MONOXIDE POISONING that the actual cause of death was broncho-pneumonia due to an infecting organism which produced nitrite. These authors state that the double-banded spectrum of NO-haemoglobin is similar to that of CO-haemoglobin and oxyhaemoglobin except that the bands are much less sharply defined than those of oxyhaemoglobin and somewhat less sharply than those of CO-haemoglobin ; the NO-haemoglobin band in the yellow extends also to a slight distance on the red side of the D line.The positions of the alpha and beta bands of the three haemoglobin compounds under discussion have been determined by Hartridge,’ and the wave-lengths of these are given below. Alpha band. Beta band. A.U. A. u. Oxyhaemoglobin . . 5768 5398 CO-haemoglobin . . . 5714 5360 NO-haemoglobin . . 5785 5418 Examination of these figures shows that there is no reason why NO-haemog,Jbin should be mistaken for CO-haemoglobin if the reversion spectroscope is employed, for if NO-haemoglobin is present the span will be a minus instead of a plus figure. 1. 2. 3. 4. 6. 6. 7. 8. 9. REFERENCES. Lewin L. Die Kohlenoxydvergi~~ung.Berlin Julius Springer 1920. Review by Alice H.M. Stationery Office 1930. Henderson Yandell; Haggard Howard W.; Teague Merwyn C.; Prince Alexander L.; and Wunderlich Ruth M. Henderson Yandell; and Haggard Howard W. New York Chemical Catalog Co. 1927 p. 110. Ibid. p. 108. Hartridge H. Hartridge H. Lancet 1928 214 1137. (ANALYST 1928 53 395.) Banham H. A. L.; Haldane J. S.; and Savage T. Hartridge H. Hamilton. J. Ind. Hyg. 1921 3 No. 2. Summary Brit. Med. J. 1930 1 No. 3607. J. Ind. Hyg. 1921 3 Nos. 3 and 4. Noxious Gases. Proc. Physiol. Soc. January 31 1920. Brit. Med. J. 1925,2 No. 3370 187. J. Physiol. 1920 54 No. 4; 1922 57 47 (ANALYST 1923 48 341); (ANALYST; 1925 50 520.) PYOC. Roy. Soc. 1923 A 102 575 (ANALYST 1923 48 351). NoTE.-A valuable summary of the literature on carbon monoxide poisoning (with a bibliography of 195 references) is given by R.R. Sayers and Sara J. Davenport in Review of Carbon Monoxide Poisoning Public Health Bulletin No. 195 Washington, D.C. United States Government Printing Office 1930. ROYAL NAVAL MEDICAL SCHOOL, ROYAL NAVAL COLLEGE GREENWICH S.E. 10. * DISCUSSION. Dr. ROCHE LYNCH said that he was very glad that Mr. Frederick had brought this instrument to the notice of the Society as he thought it was not at all well known. Most analysts had to examine blood at times and there was no doubt that this instrument provided the most suitable method of analysis. In his experience it was capable of much greater accuracy than the Haldane method. With the Hartridge instrument one could work with low percentages and deal with even 4 or 5 per cent.of carbon monoxide; this would be useful in cases of chroni FREDERICK CARBON MONOXIDE POISONING 569 carbon monoxide poisoning. There was a great deal of this about although it was not generally recognised. In many households there were gas jets gas rings, geysers etc. constantly alight and having no flues and these were all giving off carbon monoxide into rooms where people spent a great deal of time. He believed a considerable number of minor ailments were due to people inhaling carbon monoxide in this way. Mr. Frederick had pointed out that 80 per cent. saturation would cause death for certain and he (Dr. Roche Lynch) was inclined to think that probably 60 to 65 per cent.could cause death. However it must be remembered that if the patient were taken out in a moribund condition and lived for a few hours it was surprising how quickly the percentage in the blood fell and very possibly only about 15 per cent. would be found. Also, there were cases of carbon monoxide poisoning where the patient lived for two or three days and then died. Here two factors had to be considered-the factor of the percentage in the blood and the factor of the damage done to the tissues as a result of exposure to the deoxygenated blood. Regarding the figures for suicide from coal gas he remarked that this was a very easy way of terminating life and, although one would like to see an even greater tightening of the Poisons Act, the narcotic poisons available to the public were really very few and consequently coal gas helped to bring up the suicide figures considerably.Sir BERNARD SPILSBURY thanked Mr. Frederick for his observations on nitric oxide haemoglobin which tended to confirm his own opinion. He must plead guilty to having used the Haldane method. Dr. Haldane claimed to read to within 3 per cent. and he (Sir Bernard) thought that he could read to within 5 per cent. which of course was satisfactory where large amounts were con-cerned. He admitted that the test certainly required delicacy of judgment in matching the sample under examination with the standard. His investigations had always been made upon the bodies of persons found dead. In these cases the blood was difficult to examine owing to its coagulation after death and he wondered if the same difficulty were likely to arise with the Hartridge reversion spectroscope.The figures given by Henderson for a fatal result seemed to be unreasonably high. Of course here again one had to take into consideration th'e difference in the results obtained by the two methods but he himself generally found the fatal figure to be about 60 per cent.; it had reached 70 but never ap-proached 80 per c a t . The figures given werej presumably for normal people. whereas people not in normal health succumbed more rapidly. He agreed that the Hartridge reversion spectroscope was a great improvement for lower satura-tions of the blood and was very useful in cases of chronic carbon monoxide poison-ing. Quite recently a medical friend had asked him to investigate certain illness, the cause of which he could not discover.Children were always ill when living in the house although quite well when away. On examining the premises Sir Bernard had found in a basement kitchen a destructor used for getting rid of household refuse. It was a coke furnace standing near the kitchen stove with a flue from the top of the furnace to the top of the chimney. However there =as a loose-fitting lid covering the hole into which the rubbish was put and gases could very easily escape from this into the room. A mouse was placed on the kitchen mantelpiece and in three days it died death being proved to be due t o carbon monoxide poisoning. Such instances must be fairly common in ordinary house-holds and in these days of almost universal motoring also in many badly venti-lated garages.For these cases the Hartridge method of investigation would be of very great advantage. Mr. W. J. A. BUTTERFIELD said that it was very interesting to hear an account of this instrument although he had no experience of its use. The figures for th 570 FREDERICK CARBON MONOXIDE POISONING number of deaths from coal gas poisoning during 1918 and 1928 were probably fairly comparable as regards suicides but he was quite sure from a study of the statistics which he made some years ago for the purpose of a report to the Board of Trade published in 1924 that it was very misleading to take a single year’s accidents and compare them with any other single year ten years later. They varied very much from-year to year and to get a fair comparison he was sure that three consecutive years at least should be taken.For instance comparison of 1917 1918 and 1919 with 1927 1928 and 1929 would be much fairer. One got an enormous difference by picking out individual years ten years apart; if one took the next year in each case one got entirely different figures. It had been suggested that the increased number of accidents was due to the increase of the proportion of carbon monoxide in the gas. As regards 1918 and 1928 the carbon monoxide in gas in this country was much higher in 1918 than in 1928 owing to the emergency emeasures adopted during the war years to produce gas to keep up the supply. During some of these years the carbon monoxide was much higher than it had ever been since.How misleading comparisons of single years could be might be appreciated from the facts that the mean proportion of carbon monoxide in all the gas supplied in Great Britain was 15.2 per cent. in 1921 when the number of accidental deaths was 126 and 14.4 per cent. in 1922 when the accidental deaths rose to 203. Regarding the susceptibility to small proportions of carbon monoxide in the air breathed any reference to headaches resulting therefrom must be referred to the previous experience of the individual; otherwise one might be misled very much. People living in an atmosphere free from carbon monoxide were more susceptible to small doses. He gave one illustration-the drivers and stokers of the engines on the old Metropolitan Railway in the time of the steam trains were in the tunnels a considerable time inhaling carbon monoxide which (between King’s Cross and Edgware Road) incapacitated anyone who tried to walk through the tunnel quite apart from the sulphur dioxide.These men felt quite well and examination of their sick-club books showed them to be a very healthy set. The same thing was experienced in gas works and in garages. Mr. Butterfield added that he had mentioned this paper to Dr. Haldane but as the title suggested that it was simply a description of the instrument Dr. Haldane had thought it unnecessary to postpone a visit to the North in order to be present at the meeting. With regard to the treatment with carbon dioxide mixed with oxygen which was clainied to be of Arliericaii intl-oductim he did not think Prof.Henderson claimed it as originating in America. The actual origin was due to observations in war time in Europe and the theory was certainly known in this country before the practice was developed and brought into general use in America. Mr. W. PARTRIDGE said that treatment with oxygen mixed with carbon dioxide was claimed to hasten respiration. Speaking from memory he was under the impression that about ten years ago an American committee dealing with the proportioii of zarboii riioiioxidz iii t h e streets had f ~ u n d that t h e additim of carbon dioxide reduced the toxicity of carbon monoxide. The years 1918 and 1928 were bad ones to be taken for accidental deaths because the peak year was about 1922 or 1923 and at the later date he believed the companies were using drawn pipes and not seamed tubing for gas.Dr. Roche Lynch had mentioned that fewer poisons were available in 1928 than in 1918. He (Mr. Partridge) did not think that many of the additional poisons scheduled as dangerous in the ten years mentioned were generally available to the public. Mr. FREDERICK replying said that there was little he need say in reply, since the remarks made were not so much in the nature of criticism as valuable . FREDERICK CARBON MONOXIDE POISONING 57 1 contributions to our knowledge of the subject of the paper. In any discussion of poisoning by carbon monoxide there was a tendency for undue importance to be attached to the question of suicide by this means; suicide statistics were influenced by so many factors that he preferred not to pursue this aspect of the subject.He was sorry that Dr. Haldane had not been present as his authoritative remarks would have added still further to the interest of the discussion. With regard to the treatment of carbon monoxide poisoning by a mixture of oxygen and carbon dioxide this was really outside the scope of the paper and had only been mentioned as a matter of general interest. Communications on the foregoing paper :-THE INVESTIGATION OF CARBON MONOXIDE POISONING. By J. S. HALDANE C.H. F.R.S. As Mr. Frederick's paper conveys the impression that in the examination of blood from suspected cases of carbon monoxide poisoning the use of spectroscopic methods and particularly of the Hartridge reversion spectroscope is essential, I should like to point out that this is not the case.The colorimetric method as applied to diluted blood requires no special apparatus,' can be made with a single drop of the blood and when it is made quantitative gives very exact information as to the percentage saturation of the haemoglobin with carbon monoxide. I do not think that an analyst requires to use a more complicated method; it is this simple method which I have used exclusively in the numerous investigations which I have made of carbon monoxide poisoning or in which carbon monoxide was used as a physiological reagent. In cases of suspected nitrite poisoning the colorimetric method gives information at once if the blood sample is taken during life when much methaemoglobin is present; and after death the presence of nitric oxide haemoglobin is a t once revealed by the colour of the clot formed on boiling.THE PREPARATION OF NITRIC OXIDE HAEMOGLOBIN. By H. HARTRIDGE M.D. F.R.S. Nitric oxide haemoglobin was prepared by Gamgee by passing nitric oxide gas through blood or a solution of haemoglobin. Since nitric oxide reacts with oxygen to form nitrogen peroxide and since this reacts with water to form nitric acid which would change the haemoglobin to acid haematin oxygen must be excluded during the reaction of the nitric oxide with the blood or haemoglobin, or alkali must be added previous to the reaction so as to neutralise the nitric acid which will be formed. In order to exclude oxygen a stream of neutral gas such as nitrogen or hydrogen may be passed through the apparatus in which the nitric oxide is to be generated and through the blood which is to react with it.If this neutral gas is oxygen-free the oxygen in combination with the haemoglobin may be dissociated from it by warming and shaking the haemoglobin during the passage of the neutral gas. Hufner prepared nitric oxide haemoglobin by first passing pure carbon monoxide gas through the apparatus and the blood. This gas displaced the oxygen both from the apparatus and the blood with the formation of carbon monoxide haemoglobin. This compound on passing the nitric oxide gas was then decomposed into nitric oxide haemoglobin with the liberation of the carbon monoxide Nitric oxide haemoglobin can be more readily obtained by addin 572 FREDERICK CARBON MONOXIDE POISONING to blood a solution of a nitrite of an alkali metal and a suitable reducing agent.Dilute solutions of sodium nitrite and ammonium sulphide are commonly used. THE PROPERTIES O F NITRIC OXIDE HAEMOGLOB1N.-It iS a red crystalline compound (Hermann). Its solutions are less orange than those of oxyhaemoglobin, but are more orange than those of carbon monoxide haemoglobin. Two absorption bands are present in the visible spectrum which superficially resemble the oxy-haemoglobin bands in position but are more diffuse. A comparison of the position of its absorption bands with those of oxyhaemoglobin and carbon monoxide haemoglobin with the reversion spectroscope shows that they are on the long wave-length side of those of oxyhaemoglobin; that is on the opposite side to those of carbon monoxide haemoglobin (see Table at the end of Mr. Frederick's paper). Experiments on the stability of nitric oxide haemoglobin show that whereas nitric oxide can displace carbon monoxide from its combination with haemoglobin the compound thus produced is not so stable but tends to change spontaneously into methaemoglobin. On heating a ' solution of carbon monoxide haemoglobin, coagulation commences at about 65" C. whereas on heating a solution of nitric oxide haemoglobin to 50" C. it spontaneously changes to alkaline methaemoglobin, which then coagulates