326 PROCEEDINGS OF THE ACADEMY OF [1887. 



Aug. 1st. and placed them in a glass of water in my library, and 

 watched their behavior. 



The unopened flower, at 7 P. M., is very interesting from a small 

 tooth bearing a black gland, seemingly situated in the axis between 

 the lobes. It ia really on the left hand side of each lobe, and gives 

 to each division when carefully separated the outline of a mitten, 

 where in covering the hand the thumb only is free. At 8 P. M. 

 the lobes have parted, and what one would take to be a pair of uni- 

 ted sagittate anthers, covered with pollen, have advanced just their 

 length above the corolla-lobes. At 8'5, this length is doubled. The 

 same progression continued till 8'30 when growth ceased, the slender 

 "filament" having grown six lines in thirty minutes. 



Examining next the four anthers in the open flower, they were found 

 apparently destitute of pollen, and, remembei'ing that systematists had 

 found a close relationship between Rubiacece and Com2JOslfce; the idea 

 suggested itself that the stamens in the corolla were sterile, and that 

 in some unaccountable manner two stamens had become as entirely 

 consolidated with the style as in OrcJiidece, and that only two fer- 

 tile anthers were left, which had united and formed a cap wholly 

 covering the stigma. Acting on the suggestion a pair of fine tweezers 

 were placed under the seeming anthers, and a gentle lift took it oflT, 

 as if it were a thimble in miniature, leaving a clear greyish-white 

 ovate stigma behind. The sagittate form of an anther is so plain, 

 and the four angles that these would make united back to back so 

 apparent, that I am satisfied no one, at first thought, would take them 

 for anything else than as described. Full of enthusiasm over the 

 mystery I sent some to my friend Mr. Sereno Watson, to learn what 

 he would think of the "missing link" in this pair of monodelphous 

 anthers. His reply that he found some grains of pollen in the true 

 anther sacs, and only pollen on the stigma, led me to look into the 

 matter again, and I found he was right. The four anthers mature 

 before the pistil takes its rapid start. At anthesis the anthers are 

 pressed firmly over the stigma. When the growth of the pistil oc- 

 curs, the stigma wipes out, almost clean, the entire mass of pollen, 

 and so nicely as to retain the form of the anther lobes on the stigma 

 as the style develops. I have since found that this simulation of the 

 form of real anthers is not seen in the open air. The motion of the 

 atmosphere or possibly the jar from the visits of nocturnal insects 

 gives a rounded form to the stigma-covering mass. 



