570 



is distended where tlic frost tear opens internally and thus loses a part of 

 its elasticity. 1 f the tree later grows thicker, the bark exercises a lesser 

 pressure on the cambium at this place and the additional growth is. there- 

 fore, locally increased. Outwardly the trunk is not round but has ridge- 

 like processes." 



I assume a \ery similar course in the production of the structures, 

 which I term frost blisters. 'J'hese are broadly conical, but usualh flattened 

 processes at times one centimeter high on the smooth bark of trunks or 

 branches two or more years old. 



These blisters should not be confused with the conical bosses, occur- 

 ring not infrequently on luxuriantly cultivated varieties. In them a hani 

 ivood core is recognized immediately under the bark, while the frost wood 

 blister, in some cases, consists permanently of a soft tissue mass, easily 

 indented with the finger nail which, in other cases, lasts only during the 

 year of its production. 



The projections, strongly lignified from the start, and for which I 

 would like to propose the name of "duct boss," almost always have a 

 definite position in relation to the bud, while the frost blisters are found on 

 any part of the young trunk or the branch internode. "Duct bosses" are 

 bark covered, wood swellings with one or two points, which, like the begin- 

 nings of a gnarl, i)roject above the perijjhery of the rest of the wood body. 

 They owe their production to the excessi\e develoi)ment of the two \ascu- 

 lar bundles which normally pass into the bud cushions and unite with the 

 central, strongest bundle of the vascular bundle body of the petiole. 



In the tender frost blisters we find no connection with the cords of the 

 leaf spur. They are found at any place and arise from a blister-like dis- 

 tention of the bark body away from the wood cylinder. The young wood, 

 lying on the wood cylinder, at once begins cell increase, since the distention 

 takes place only in late frosts and, therefore, at the time of growth activity. 

 This young wood fills the ca\ity with a thin-walled parenchyma wood 

 which gradually [)asses over, at the j)eriphery, into normal wood. 



The wdiole process taking place here is the same as occurs in the new 

 formation of bark on an artificially produced wound surface. In blister 

 formation the dift'erence lies alone in the fact that the bark does not scale 

 off, but is only raised in places by the frost and that thereby the new struc- 

 tures, lying above the wood body, at first do not become visible to the 

 naked eye. At times they can be very clearly recognized by their unusual 

 luxuriance when the bark is cut on large frost blisters. It is then possible 

 to expose here and there a wrinkled outgrowth, several centimeters long 

 and 0.5 to T.o cm. high, which is not connected at all with the old bark and 

 only rests on the wood body. In one case (in the pear Bonne Louise 

 d' Avranche) the outgrowth had ruptured the bark mantel and had ex- 

 tended far beyond the circumference of the trunk as an irregularly outlined, 

 somev^'hat conical mass with a wartv, crumblv surface. 



