342 EVENINGS AT THE MICEOSCOPE. 



and the result is that there are three oval apertures. 

 The next step is that on the top of the four united 

 globes two other dumb-bells stand erect, and lean over 

 towards each other till their upper globes come into 

 contact, their lower ones remaining remote ; these are 

 soldered to the mass and to each other, at the points of 

 contact, leaving a fourth aperture. The same is re- 

 peated at the opposite end by two other dumb-bells ; — 

 and the structure is complete as you see it. In almost 

 all cases the two united globes of these terminal ele- 

 ments are fused into one 

 globe, and in not a few 

 instances the appearance 

 is as if these two dumb- 

 bells were but one, bent 

 over in a semi-circular 

 form ; but still a good 

 many specimens occur in 



DUMB-BELLS M nOLOTHTOIA. ^J^J^jJ^ ^^B tWO dumb-bcUS 



can be quite distinguished from each other. The cal- 

 careous matter that solders the elements together seems 

 abundant, and has the appearance that would be pre- 

 sented if they had been made of solid glass, and united 

 by glass in a state of fusion ; the latter having appa- 

 rently run together, so as to smooth and round angles 

 and fill up chinks, even where, as is often the case, the 

 globes themselves have only mutually approximated, 

 and not come into actual contact. 



The average dimensions of these oval aggregations 

 may be "OOi inch in length, and a little more than -002 

 in width ; but some specimens occur which are a little 

 larger, and others a little smaller than this ; while the 

 irregular plates ai'e sometimes three times the length. 



