has felt absolutely well for a year. He has no demonstrable sign of disease about him. Case XIV. W. A. McN. He came to us with the signs of an early advancing tabes. He was chiefly troubled by pains in the legs and stomach. His mind was slow and he could no longer perform the duties of a train dispatcher. His blood Wassermann reaction was strongly positive. Unfortunately, the record of the condition of the spinal fluid, which was examined at this time, is not available. Under mercury he made very slight improvement. He has received in all 16 intravenous injections of salvarsan or neosalvarsan, equivalent to 8.75 Gm. of salvarsan. The blood Wassermann reaction has been negative for nearly a year. The spinal fluid, which was examined 10 months ago, showed no increase in globulin, 3 cells to the cumm., and a Wassermann reaction negative in 0.5 cc. but strongly positive in 1.0 cc. He has been back at his old post, which is one of great responsibility, for a year. He has gained much weight. His lightning pains have gone. He is steady on his feet. His pupils, which were unequal and irregular when we first saw him, are now equal and regular. The knee jerks remain absent. As far as his work, his life and his efficiency are concerned he is a well man. On the other hand, there are cases in which intravenous injections of salvarsan seem to be insufficient. In such cases, while the clinical improvement is often marked, the serological symptoms of the disease remain practically unaffected. The following instances are typical of this class of cases: The first instance was a man of 49, who had had indefinite symptoms referable to the central nervous system for seven years. He has been unsteady on his feet for five years and has had ever-increasing numbness and tingling in the feet, accompanied by severe attacks of pain in the legs, and very troublesome headaches. He recently found that his mental acuteness and ability to work was much affected. His pupils were unequal and irregular and did not react to light. His knee and ankle jerks were absent. There was incoordination and marked sensory changes. In addition to this, he carried a systolic blood pressure of 190 with a diastolic pressure of 140. There was a faint trace of albumin and a few casts in the urine. The blood Wassermann reaction was negative but the spinal fluid showed 124 cells per cumm., and a Wassermann reaction positive in 0.2 cc. He has received 6 intravenous injections of salvarsan, a total of 2.3 Gm. His spinal fluid now shows a faint trace of globulin and 25 cells per cumm., but the Wassermann reaction is quite as strongly positive as it was before treatment was begun. While the physical signs have changed scarcely at all, the symptomatic improvement has been very marked. The pains and headaches have gone. The numbness and tingling are greatly lessened and the ability to work at high pressure has returned. It is interesting to note that the patient has had no untoward effects from the injection of salvarsan despite the high blood pressure and evidences of nephritis. While the symptomatic improvement has been very marked, the Wassermann reaction in the spinal fluid is yet unaffected, and it is likely that it will not be much changed unless the intraspinal treatment is administered. A second instance of this kind is that of a man 37 years old, who was infected with syphilis 16 years ago, for which he received very little treatment. He had been well up to two weeks before we saw him, when he began to have terrific spells of nausea and vomiting associated with dizziness. The pupils were unequal and irregular and did not react to light. There was no sensory disturbances. The blood Wassermann reaction was strongly positive. The spinal fluid was not examined at this time. The patient received intramuscular injections of mercury for 4 months, under which the blood Wassermann reaction became negative and there was an abatement in his symptoms. He then received 6 intravenous injections of salvarsan or neosalvarsan, an amount equivalent to 3.2 Gm. salvarsan. During this time the symptoms entirely cleared up and the patient felt as well as he ever did. He not only worked hard but played baseball and labored in his garden. The spinal fluid examined after the 6th injection of salvarsan showed a marked globulin increase, 163 cells to the cumm., and a Wassermann reaction positive in 0.2 cc. In this case we still have signs of a very active syphilis of the central nervous system which at present is not causing symptoms. It is probable that this condition cannot be controlled satisfactorily without the intraspinal administration of salvarsan. In conclusion, we wish to state our opinion in regard to the treatment of syphilis of the central nervous system. (1) There are undoubtedly cases in which an arrest of the process may be accomplished by the use of mercury alone. On the other hand, there are a very definite number of cases of syphilis of the central nervous system in which the action of mercury is positively detrimental. In these cases, salvarsan frequently acts extremely well and is an invaluable addition to our armamentarium. (2) There are many instances in which the process may be arrested by the intravenous injection of salvarsan alone. These are usually meningo-vascular involvements, and the best results are to be expected in the early cases, especially those which have begun to have symptoms, but in which the physical signs of the disease are either absent or slight. Whether the abolition of all the serological signs of the disease can be expected under the intravenous administration of salvarsan is, as far as our experience goes, still open to doubt. (3) When mercury given intramuscularly or the intravenous injections of salvarsan have failed to arrest the process satisfactorily, the intraspinal administration of auto-salvarsanized serum will have a markedly beneficial effect in a large proportion of cases. (4) The form of affection most amenable to treatment is that with meningo-vascular involvement. The greater the central degeneration the less can be expected from treatment, for there is no drug which can re-establish nerve tissue once it has been destroyed. As it is often difficult to differentiate clinically between these types of the infection, active and intensive treatment should be tried in most cases. However, it is probable that very little can be done in most cases of general paresis. (5) The Swift-Ellis method for the intraspinal administration of auto-salvarsanized serum, when carefully administered and rigidly adhered to, is a safe and eminently proper procedure for the combating of syphilis of the central nervous system. Bibliography 1. Adler, H. M.: Boston Med. & Surg. J. V., CIXXI, No. 24, p. 5. Swift, H. G., and Ellis, A. W. M.: N. Y. Med. Jour., 1912, XCVI, Ravaut, P.: Bull. et men Sec. med. d nop d Paris, 1913, XXXVI, p. 53. 6. p. 752. 7. Wile, U. J.: J. A. M. A., April 11, 1914, p. 1165 J. A. M. A., July 11, 1914, p. 137. 8. Ogilvie, H. S.: J. A. M. A., Nov. 28, 1914, p. 1936. 9. AN ANALYSIS OF THE DEATH CERTIFICATES By G. E. HARMON, M. D., C. P. H. Department of Hygiene, Western Reserve University, and Bureau of Vital Statistics, Division of Health, Cleveland City, State and Federal bureaus dealing with vital statistics are endeavoring in every way possible to improve and to increase the accuracy of such statistics. The idea is to make the data relating to vital phenomena gathered by statistical methods as accurate as possible, so that reliable deductions may be made and readily available information obtained which shall be of real value to the movement to procure more healthful conditions. It is also extremely important that uniform methods be used in compiling the data of vital statistics, so that valuable comparisons can be made between different cities, States, etc. To attain these ends with accuracy and uniformity, the Health Department endeavors to secure correct certification of the cause of death on each certificate filed, and uses the International List of Causes of Death in classifying deaths. In order that the International List of Causes of Death may be used, it is essential that the causes of death on the certificates should be definite and accurate. Ill-defined and loose terms should as far as possible be eliminated, for otherwise the death cannot be assigned to its proper place in the classification. To obtain this much-to-be-desired accuracy, which is essential to efficient vital statistics, the Health Office returns regularly a number of death certificates to physicians in order that corrections or additions may be made in the medical certificate of death or in order that the cause of death may be more definitely stated. The success of this plan depends upon the assistance and co-operation of the practicing physicians. It is to be hoped that all physicians in making out death certificates will give the cause of death as accurately and definitely as possible. It is also hoped that when asked they will co-operate in making any alterations or corrections in the way of greater clearness by attending to the matter at once and returning the certificate at the earliest possible moment. All these details will greatly aid in improving our vital statistics. This paper will be devoted to an analysis of the death certificates which were returned to physicians for correction from January 1 to June 30, 1915. As far as possible the reasons for returning the certificates will be discussed and explained. It is hoped that this discussion will be of some assistance in showing what terms should be avoided in returning deaths, and also that it will be of some help to physicians in the task of giving the causes of death more accurately. It will be necessary to make frequent reference to the International List of Causes of Death in order to bring out some of the points. A consideration of the reasons for returning the certificates will be taken up in the order of their importance, beginning with the reason for returning the largest number of certificates. During the period under consideration, 4,749 certificates were registered at the Health Office, and of this number 422 or 8.9 per cent were returned for corrections. Seventy certificates were returned because the type of pneumonia was not stated. In general two types of pneumonia are recognized, the broncho and the lobar. In order that the deaths caused by pneumonia may be correctly classified, it is necessary that as far as possible the type of pneumonia be definitely stated as of the broncho or lobar type. It is not sufficient merely to state pneumonia as a cause of death. If the type cannot be determined the proper procedure is so to state on the certificate. If the pneumonia is secondary to some other disease, such as typhoid fever, whooping cough, measles, phthisis, etc., the primary condition should always be stated in addition to indicating definitely the type of the pneumonia. Seventy-nine certificates were returned on which the cause of death was given as due to external causes or violence. Very frequently it was not stated whether the death was due to accident, suicide or homicide. It is necessary that a definite statement be made as to whether the deaths were due to accident, suicide or homicide, since the International List makes provision for these three classes of deaths due to external causes. In addition the means or instrument of death should be indicated in each case. This information was omitted on many of the certificates and so they were returned in the endeavor to obtain it. So in every case of death due to violence the nature of the |