(3) these dwarfed systems do not flower. Such may occur in Abies pectinata in response 

 to attack of the fungus Exoascus, also occasional in Pinus sylvestris. 



(2) Chermes viridis, an Aphis, is responsible for the curious Pine-apple gall of 

 the common Spruce (similar galls on Sitka Spruce). All the foliage- leaves of the 

 Picea annual shoot are laid down in June of the preceding year, in a normal 

 Fibonacci sequence (5 : 8), and are protected by a dense investment of bud-scales 

 over the summer and ensuing winter. Hibernating females of Chermes attack the buds 

 as soon as the hibernaculum of bud-scales loosens in spring (April), commonly 

 unilaterally, penetrating to the apical cluster of leaf-primordia, boring the tissues, and 

 producing a numerous progeny similarly feeding on cell-sap. In response to the 

 primary stimulus the leaf-units more or less cease further differentiation, and reduce to 

 mere spinous processes, while the leaf-base develops as a peltate scale in the manner 

 of the cone-scales of Cupressineae. The axis enlarges, the intercalated leaf-base 

 growths are forced up laterally, leaving open pockets in the leaf-axils, and the tissues 

 fill up with large (20 //.) starch-grains. 



Following the existing phyllotaxis-construction, the enlarged leaf-units present 

 a compact cone-like aggregation; but contact-relations may be reduced to parastichies, 

 8 : 13, as the leaf-bases subtend a low angle. The intercalated scale-growths make 

 rhomboidal facetted areas, to 7 mm. broad, and such a cone-like construction may 

 include 50-70 leaf-units in spiral series, wholly centric, or on one side only of the 

 shoot-axis ; all intermediate stages may be found between complete arrest of the entire 

 system and the affected region being unilateral, or with fully proliferating apical section. 

 When fully grown the whole may be 40 mm. long and 16 mm. diam. (A second 

 species \C. Abietis\ giving pale greenish galls, -|-in. diam. or less, on similar construction- 

 principles, is also common, and the two forms may occur on adjacent shoots of the 

 same Spruce.) 



While these growths are taking place the larvae wander in and out of the cavities, 

 but the edges of contact are ultimately sealed by interlocking papillae. Larvae 

 enclosed in the gall-loculi develop pupae ; those outside die off, and many pupae (to 50) 

 may be found in one chamber. The contact-lines of the rhomboidal facets are 

 emphasized on the axillary side of the leaf-units by papillose growths, which may be 

 brightly coloured with anthocyan in the cell-sap. The pupa stage is attained in June, 

 and from mid-June to July the perfect winged insects escape, often in great profusion. 

 Starch disappears as the tissues sclerose and dry up ; the loculi gape widely, opening up 

 the cavities to the exterior, and the residual dry, dead, brown galls may persist on the 

 shoots for 2-3 seasons. 



Note. The growths once initiated are formed whether they contain imprisoned 

 aphides or not. The opening mechanism is also provided by the plant as the new 

 leafy shoots attain adult efficiency and drain the older portions of food and water in 

 the hot season (end of June). The inner tissues become sclerosed and pitted, starch 

 is depleted, and with the loss of water the tissues die. The peltate-scale effect, 

 obvious at first on cutting a section down one of the steeper parastichy lines, is wholly 

 lost as the pockets shrivel back and gape open. 



Botanical interest centres in the fact that the shoot-system, under stimulus of 

 attack by the parent aphis, repeats certain elementary factors of cone-formation no 

 longer noticed in the Picea ovulate cone, but directly parallel with the case of Biota 

 in the Cupressineae ; e. g. in succulence, intercalary extension below the arrested leaf- 

 primordium, and more especially in the sealing mechanism of papillose hairs. The 

 loculi of the cone-formation are to all intents identical with those formed to protect 

 the developing seeds of the Cupressoid cone. In this way vestigial factors of cone- 

 elaboration may appear in the history of a race in response to a new stimulus, and 

 this happy accident enables the Chermes larvae to obtain protection and easily taken 

 food in early spring. Hence the second summer brood, produced at a time when 

 there are no breaking buds on the Spruce, cannot form galls, and live exposed, or even 

 migrate to other Conifers (Larix \C. viridis], Pseudotsuga, Pinus sylvestris) ; but no 

 other Conifer produces these growths. The fact that the ' galls ' induced by different 

 species of Chermes are different in details as size, colour, degree of closing, indicates 

 that the stimulus must be due to some excitant material ejected by the fundatrix, and 

 that this may be variable, hence coming under the operation of natural selection 



27 



