35 



The contraction of the cement from the wood leaves a space 

 which is likely to become filled with water. In winter the space 

 may open like a frost crack and decay set in in time, even if the 

 wood is disinfected ; therefore, we have always found it important to 

 have the cavities painted with some thick substance of an elastic 

 nature to fill the space caused by the contraction of the cement. In 

 this way the water is kept away from the wood and the danger of 

 infection lessened. 



Various kinds of fillings have been tried, also different substances 

 for disinfecting cavities, but little can be 

 said about their relative value at the pres- 

 ent time. The writer has experimented 

 with a variety of cements and disinfectants 

 and prefers an oily substance to a watery 

 solution for disinfecting a cavity. We 

 have used creosote more than any other 

 substance for this purpose, since its wood 

 preserving qualities are well known, and 

 so far as we l^ve seen it causes no injury 

 when applied to cavities. 



One of the objections to the use of Port- 

 land cement in filling cavities, as previously 

 mentioned, is its tendency to crack. Mr. 

 John T. Withers, who has hgid much ex- 

 perience in filling trees, recommends re- 

 inforcing the surface with wire. He has 

 employed this method, and some of our 

 best fillings have been re-inforced in this 

 way. 



The fillings in trees which are likely to 

 sway considerably are sure to crack, and 



although the cracks formed are not often serious, this objectionable 

 feature may be overcome by laying the cement in blocks or sections 

 and filling in between each section with some thick elastic substance 

 to keep out the water. Such substances as thick coal tar, coal tar 

 and roofing tar mixed, or slaters' cement would answer this pur- 

 pose. 



Fig. 15. Demonstrating the 

 object of treating cavities. 

 Upper figure showing cavity 

 of long standing, with callus 

 curved in, which, if it had 

 been filled, would be as repre- 

 sented below. 



