Since 1994, our clinic has had monthly study meetings for those who looked after bedridden patients in their homes. As of June 1996, we have gotten together 39 times and a total of 168 persons have participated (mean 4.8). We have discussed the welfare system (3 times) and diseases such as bed sore and lumbago (16 times), visited participants' homes and demonstrated medical equipment (4 times each) and talked about general affairs (12 times). In home health care, a role of those who look after bedridden patients in their homes is very important. Social and medical systems should support them. We think ther meetings could continue to support them morally and psyehologically.

We have provided home health care to 42 patients with terminal cancer over the past five years. The pancreas was the most common site of the primary cencers (9 cases). The average age of the dead was 74.6 years, which was younger than that from heart failure or brain infarction. The average duration of stay of home was 60.1 days. The average frequency of visit to a patient's home was 15.8 times. In 52.4% of the cases, morphine were administered for pain control. The average dose was 48.2 mg per day, and term of administration was 30.6 days. In most cases, those who looked after the patients were daugh-in-laws or wives.
In some cases, a local government lent a bed to a patient. Of the patients 21.4% were fold by their physicians what they are up to were. In conclusion, home health care of cases of terminal cancer will be more important in the future. Pain control by morphine, support for those who look after the patients by frequent visits, practical use of the welfare system and full-time medical care system which meets patients' need are thought to be necessary.

A number of investigators have been concerned with the anthropometric as well as physical characteristics of sumo wrestlers during the past several decades. However, none have attempted to determine body density and percent body fat by the most accurate technique of hydrostatic or underwater weighings. Thus, a precise quantification of body density and percent body fat for the wrestlers has not yet been obtained. Ogawa et al. (1972) have predicted the body composition of sumo wrestlers from a formula which added the triceps and subscapular skinfold thicknesses. As the regression equation used in this instance was based on samples of normal young men, it may not have great predictive accuracy when used on the exceptional population of very stout athletes.
The purposes of this study were : (1) to assess the validity of percent body fat estimation by skinfold thickness measurement when compared to the hydrostatic weighing criterion method ; (2) to determine the interrelationships between anthropometric variables for use in the clear assessment of the physical characteristics of ‘Shindeshi’ sumo wrestlers ; and (3) to thereby develop a formula that would reliably predict percent body fat using skinfold thicknesses and/or anthropometric measurements in this exceptional population. The interrelationships between anthropometric and body composition variables were investigated using 35 Shindeshi sumo wrestlers, aged 15-20 years (X=16.5±1.5) .
The results of the present study can be summarized as follows.
1. The Shindeshi in the present study possessed physiques more developed than those tested in the past, as evidenced by Rohrer's and Ponderal Indices which were approximately 180 and 26.2, respectively. The development of body weight was particularly notable (i.e., over 100kg) .
2. Percent body fat for the Shindeshi was remarkably higher than that of a normal population of the same age, with approximately seventy percent of the Shindeshi possessing more than 20% body fat and less than 1.05000 body density.
3. Body density and height correlated negatively and insignificantly with almost all the variables. In contrast, correlations of body weight, circumferences, and all other variables were, in most cases, high and positive, with body weight correlating least with height. The skinfold measurement correlating best with % body fat was that taken at the thigh site (r=0.898) . The circumference measurements correlating best with % body fat were obtained at the thigh (r=0.888 and 0.831) and the abdomen (r=0 885) parts.
4. Of the six formulae for estimating % body fat, Sloan's correlated best with % body fat as determined by densitometry, and best approximated the mean value of 24.5%. However, all of the equations underestimated the measured % body fat of the Shindeshi.
5. When an exceptional population is being investigated, estimation of % body fat should be done with a population specific equation to ensure predictive accuracy. The following multiple regression equation (r=0.963) should, therefore, be utilized for estimation of % body fat in sumo wrestlers.
Y=0.2488x₁+0.6172x₂-14.3962 where x₁ and x₂ are abdomen circumference and skinfold thickness at thigh, respectively.