10 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



species, Calathinus and Triandrus, put into separate classes, an 

 error which admits of correction certainly, but for the present pur- 

 pose serves to illustrate the problem before us. I have myself 

 classified these plants according to the number of flowers produced 

 on a stem, and the system is absolutely perfect, until — yes, until it 

 breaks down in the presence of a plant that converts all my wisdom 

 into foolishness by producing one or two flowers more than it ought. 

 We know just enough to be kept in fear that Nature will some day 

 shatter all our plans by putting the angles where the curves ought 

 to be, and causing two and two to count as five. If we attempt to 

 classify by colour, in order to evade the difficulties arising out of 

 form, we are no better off. Take, for example, the Poet's Narciss, 

 which should be single-flowered and white, with a sanguineous cen- 

 tre. What shall we say of Bicolor, which has two flowers and a 

 yellow centre ? And, again, what shall we say of Gracilis, which 

 has two or even three flowers, and is wholly yellow ? These two un- 

 manageable beauties are so nearly allied to Poeticus that separation, 

 which at the first blush appears easy, proves at last to be impossi- 

 ble. If the varieties in their delicate graditions bring the species 

 near together, and thus perplex us, we are not the less perplexed 

 by the variations of the varieties themselves. Take the double 

 Telamonius for example. It is sometimes destitute of a corona, 

 forming a closely-packed rosette ; at other times it is double within 

 the tube only, and occasionally it shows a perfect single flower. In 

 its best state, when newly imported, it demonstrates the capabili- 

 ties of its native Italian clime. In its worst state, as an old- 

 established garden flower, it gives striking indications of the 

 deficiencies of the subarctic clime to which it has been transported. 

 There is a diminutive variety of the Poet's JSarciss called Verba- 

 nensis. It was known to Parkinson, and is by him described at 

 page 87 of the " Paradisus." He had not the least suspicion of its 

 relation to poeticus, a relation discovered two hundred years after- 

 wards, and you will find it properly placed at page 318 of the noble 

 essay on Amaryllids of the Honourable and Reverend Dean Herbert. 

 But this example is full of encouragement, because, if at the end 

 of every two centuries a difficulty is disposed of, the whole cata- 

 logue of difficulties, which for the sake of argument we will estimate 



