This section is from the book "Handbook Of Hardy Trees, Shrubs, And Herbaceous Plants", by W. Botting Hemsley. Also available from Amazon: Handbook of hardy trees, shrubs, and herbaceous plants.
The characters upon which systematic classification is founded, reside chiefly in the various modifications of the organs of reproduction and the floral envelopes. Distinguishing the organs of a plant into two sets - those concerned in its reproduction, and those that perform its nutrition - we expect in a genus some material recognisable difference in the former, or, in other words, we put together in one and the same genus all the species known which have the different parts of their flowers constructed and arranged upon the same plan; and when there are constant differences between plants which have the same plan of structure we say that these latter are distinct species. Apply this to such a clearly marked genus as Rosa or Lilium, and the application of this plan will be clearly seen. Perhaps nothing is more artificial, in a sense, than the so-called natural system of botanists. Scarcely two botanists agree as to what should constitute an order, a genus, or a species. These differences of opinion often appear greater than what they really are, for they all resolve themselves into the question of the value to be attached to certain characters. The cultivation of plants and comparisons of the same species from different parts of the world, have taught us that variability, more or less rapid or wide according to conditions and circumstances, is a prominent feature of most species under observation. What the limits of this variability are, nobody has yet determined, and some declare it to be illimitable. But this is not the place to discuss the stability of species; suffice it to say that for all practical purposes there is little difficulty. With the horticulturist it becomes a question whether a certain plant, whatever rank we may assign to it, be worthy of cultivation, either for its use or beauty. And this point decided, there is little to prevent him from ascertaining whether it will be better to propagate it direct from seed, or by some non-sexual means, as from cuttings, grafting, etc. Of course the method adopted will depend upon the easiest way of transmitting it pure.
There are no general rules by which botanists are guided in defining species. In some groups of plants, certain organs appear to be so constant in their form, number, hairiness, etc., as to characterise species, whilst the same set of organs in another group of plants vary so much as to be of no use in distinguishing species, and consequently the botanist has recourse to a different set of organs, affording more permanent and reliable characters. This, coupled with the variation of plants under diverse conditions, will explain the difficulties experienced in determining species from written descriptions. Indeed, it may safely be averred that the most accomplished botanists often fail, after careful study, to identify a plant with its description, even when that description is as perfect as it is possible to make it from half-a-dozen or more specimens; and it is usually considered necessary to compare the new specimen with the original in the case of little known species. We make this statement here simply for the purpose of warning beginners against depending too much upon descriptions, without collateral knowledge, and against being hasty in their decisions. It is hardly necessary to add that no person can expect to become acquainted with the names of plants by the use of books alone. There must be preliminary practical knowledge to render the study of botanical works profitable. As a rule we learn the names, and little more, of a great many plants, and we often go on loading our brains with these names until there is a break-down, because they are not associated with any peculiar characters possessed by their owners, but are simply based upon general appearance. Hence the need of a book of reference to refresh the memory, and assist in determining the correct name of a plant. From what we have said respecting the variability of plants, it naturally follows that all descriptions are more or less comparative in their distinctions, that is to say, without having any particular standard or starting-point, we frequently employ the words small and large, and many other terms, in describing plants of this or that group, and therefore these terms are qualified to a certain extent by the general characteristics of the plants under consideration. It is far more hazardous to give dimensions without allowing a wide margin for variation, than to limit the description to distinguishing characteristics and relative size, hairiness, etc. This course has been followed to a great extent in the present work, and it is confidently believed that it will meet with approval, and be of more service than disconnected,' though more complete descriptions.