450 Journal of the Department of Aortcultuue. 



the g'as is still stroiif^', as sliould be Ihe case with iaigo trees. The 

 residue in ilie generator slioiilcl he huried. It should not rontaln any 

 lihickt lied limi])s of cNiiiiide. The ]»ieseiice of any is iii(ii('ali\ c of 

 s(jnie fiiiill. as ilie ii>(' of loo liilh' acid. II ]iotassiuni cyanide 

 (!(fi-99 ])cr cent.) is used in place id sodium cyanide, the (|uaii1ities of 

 cyauide and (il walei- sliow n in tlic lalile >liould lit iin reined h\ oii(>- 

 lliiid. uliilc Ihe (|uanlit_\ of a( id is kepi uTicliauj^'ed . 



Basis for Calculating Dosage. — The amount of cyanide necessary 

 lo kil] a scale insect on a cloth -covered tree varies wilh quite a nuniher 

 of factors, hut two only need to he considered in fjainin"' a dosage 

 lalile for funiioTitioji. These are Ihe voluin(> of Ihe -.'iiclosed space and 

 the area id' clolli Ihiough -which ihe r^iis Avill leal<. The leakao'e facloi' 

 is highly iinpo)>iant; hut ihe raie of leidiaf^e varies <^ieally with va ly- 

 ing iittle circumstances, and ii follows that a dose just about riglii 

 for a tree of a given size on one occasion may he inuch too large 

 or too small on another occasion. At best a dosage tahle can only he 

 roughly approximate. The tahle given helow was carefully calculated 

 for use with sound covers of such cloth as is recommended in this 

 article. It provides a uniform strength of gas on the hasis of volume 

 Avith what is thouglit an ample allowance for leakage. The allowance 

 is ])roportioned to the area. The ratio of the volume to the enclosing 

 surface varies with the size of a space, becoming less as the size 

 increases, and hence the allov>ance for leakage has to be greater for 

 a small than for a large tree. The relative difference for large and 

 small trees in the area through wdiich the gas has to escape is 

 surprisingly great. For instance, the surface area of a tree four feet 

 liigh hy four feet wide is about five times as much in proportion to 

 the volume as is the case with a tree twenty feet high by twenty feet 

 wide. Tt follows that the smaller the tree to be fumigated the greater 

 is the importance of iiaA'ing the cloth as gas-tight as it is practicable 

 lohave it and the greater is the risk of a materially serious over-dose 

 or under-dose. Oil increases the gas-retaining property of a cloth, 

 but it decreases the durability of the (doth so much that oiling is 

 little practised on fumigation-covers. 



As the cost of fumigating in South Afidca lies largely in Ihe 

 outlay for the equipment and in the labour, it is generally good policy 

 to mak'e the desti'uction of the scale as ceidain as possible by giving 

 the tree as heavy a dose as it Avill stand without serious loss of foliage 

 or damage to the fruit. The killing back of tender twigs for several 

 inches and the dropping of some of the oldest leaves is tr. be expected; 

 and when this damage does not o( (ur it is advisable to increase the 

 dosage in further work. 



The common brown or soft scale, the oleandei oi- white scale, the 

 circular purple scale, ajid the false circular purple scale, all seem to 

 he as easily killed by the gas as the common red scale. The mussel 

 scale is considered more difficult to kill, a)id in California it is 

 customary to use a dosage stronger hy one-third for this species than 

 for the other common ones. ^^ot uncommonly, the mussel scale 

 dosage, called the full schedule or 100 per cent, dosage, is also used 

 for red scale. There has he(Mi very little experience in South Africa 

 in fumigating for the muss(d scale, and it may be advi'^able always to 

 use a stronger dosage for ii than is given in the labl<> Ihat fidjows, 



