IN JAVA. 



93 



anther-cap, as seen in Fig. 18 and in 19, where the anther-eap 

 is removed. 



On the conclusion of these singular movements no remains 

 of the stigma can be seen. As a rule these operations are con- 

 cluded before the full expanding of the flower, whose petals, 

 after remaining expanded for only a 

 few hours, fade, and, closing round the 

 column, exclude any intruder from dis- 

 turbing the interesting and mysterious 

 rites of nature being enacted within. 

 I have found that in some cases the 

 rostellum (the upper margin of the 

 stigma) is not invaginated down the 

 stylary canal, but retains the more 

 natural orchideal form of a broad flat pig. 22 

 floor to the anther, projecting far over 

 the stigma as seen in Fig. 22. When 

 the flower of Arundina spedosa has this 

 rare form it invariably, as far as my 

 observations enable me to speak, falls off unfertilised. The 

 pollinia also lie far back in the anther, and are entirely con- 

 cealed by the anther-case, which fits close down all round. An 

 insect, to secure the pollinia, would require to alight on the 



■AEUNDINA SPEOIOSA, 

 SHOWING THE SECOND FORM 



OP flower; E, P, as I>f 



PIG. 16 ; I, BID6E ON FLOOR 

 OP ANTHER H; K, BOUNDARY 

 OP ANTHEK-CAP. 



FIG. 23. FIG. 21. 



FIG3 23 AND 24.— EKIA SP., NEAR TO E. JAVEXSIS ; A, ANTHER-CAP, IN FIG. 23, 

 SHRIVELLED UP ; B, POLLINIA ; B^, POLLINIA SWOLLEN AFTER FALLING INTO 



stigma; d, kostelluji; e, stigma. 



margin of the rostellar platform and lift up the anther case, a 

 difficult operation, which supposing it to have successfully 

 accomplished, it might wander far to find a stigma to apply 

 the pollen so obtained to, for its own form of organs does not 

 8 



