294 TRANS. ST. LOUIS ACAD. SCIENXE. 



whilst the rest of the bismuth is still liquid and these grow 

 towards its centre. The inner liquid portion is forced up into 

 the capillary from this cause until the capillary itself becomes 

 clogged with bismuth crystals, when the bulb bursts with a 

 sharp crack. 



The only way of avoiding this is, of course, by having a large 

 bath of mercury and allowing the temperature to sink very slowly 

 indeed. As soon as an expansion is observable, it can be identi- 

 fied as being due to anomalous behavior of liquid bismuth by very 

 gently raising the temperature and reversing the process, when 

 the bismuth will contract instead of expanding. 



Not so when crystals are the cause of the expansion. It was 

 utterly impossible to check the growth of the crystals, prevent 

 further expansion, and the bursting of the bulb in a few seconds 

 of time following, even by raising the temperature ever so quick- 

 ly. The phenomena are quite different, and cannot be mistaken 

 the one for the other. 



I herewith give the records of my two most successful experi- 

 ments : 



No. 1. No. II. 



Wght. dilatom. -)- mercury at 300° C 11.570 grms. 14-742 grms. 



Wght. dilatom 7-7'- '' 9-160 " 



Wght. mercury 3-848 " 5-582 " 



Volume of 20 divisions of scale a.00102 c.c. 0.00151 c.c. 



Volume of bulb and capillary 0.29S87 c.c. 0.43355 '^ ^- 



Wght. bismuth in dilatometer 2.814 grms. 4.0S2 grms. 



Specific gravity bismuth at 280° 9-415+ 9*4i5 — 



Temperature of maximum density of bismuth. ...near 268° C. near 270° C. 



(Not readily and sharply determinable.) 

 Am't of rise of bismuth from maximum den- 

 sity to bursting of bulb \ div. \ div. 



Temperature bursting of bulb 260° C. 261° C. 



Evidently the point of maximum density of molten bismuth 

 lies very near 270° C. The character of the observation and the 

 high temperature jDlace a more exact determination out of ques- 

 tion. On further decrease of temperature it expands, presenting 

 anomalous densities, and this expansion is considerably less than 

 for water, so that its anomalous behavior is not so pronounced 



X 



