192 CARNEGIE INSTITUTION 



I find physicists most positively of the opinion that the walls of 

 laboratories should be broken by as few openings as possible. No 

 flues for ventilation or anj^ other purpose should be included in the 

 thickness of the walls. Plumbing and piping should be placed in 

 wells reserv^ed for the purpose and conveniently accessible. 



As material for the construction of laboratories, nothing seems 

 preferable to good brick well laid. Experiments long ago made by 

 officers of the Coast and Geodetic Survey show that sandstone should 

 be rigidly excluded from every portion of the building. I know of 

 no objection to concrete ; provided, however, that the stone used be 

 sharply angular and on no account consist of rounded!; pebbles. 



The walls of a laboratory should be very thick and massive, not 

 less than three feet in the lower story. All doors and windows 

 should be double. It is the habit to inclose the constant-tempera- 

 ture rooms in double walls, but I believe that a single wall sur- 

 rounded by cork brick would be equally good. This admirable 

 material is much used as a non-conductor ; for instance, in the in- 

 sulation of the clocks of the Paris observatory. 



Subdivisibility , — An important feature of laboratory construction 

 is the subdivisibility of the work rooms. As investigations succeed 

 one another, rooms of different dimensions become desirable. The 

 smallest room requires one window and one door, with independent 

 gas, water, and electrical connections. So far as possible, it should 

 be arranged that such rooms may be separated from one another by 

 partitions of hollow brick, which can be removed without interfering 

 with the structure of the building. Thus the size of the work rooms 

 can be reduced to a minimum or enlarged as required at a trifling 

 expense. 



Bascme?tt Work Rooms. — The basement of a laboratory should be 

 of sufficient height to be used conveniently as working space. In a 

 well drained locality it affords the very best of working rooms. 



hiterior Work Rooms. — Opinions differ somewhat as to the best gen- 

 eral plan for a laboratory. A few physicists object to internal rooms 

 lighted only from above, considering them gloomy and subject to 

 accident from the breaking of skj'lights. On the other hand, such 

 rooms afford the best constant-temperature spaces and are freest 

 from mechanical disturbance. Again, so-called wdre glass and other 

 simple expedients give ample protection from danger of falling 

 glass. For these reasons a majority of physicists approve of inter- 

 nal rooms surrounded by a corridor, and this again inclosed by outer 

 working rooms. The corridor, among other advantages, affords 



