Two years ago Mr. Airey published some interesting facts about the dust collected on smeared plates of glass, and mentioned that some of the dust was pollen. I began accordingly to enquire into the forms of pollen likely to be found in such places, and from this went on to discover all that I could about pollen. I had many years ago drawn a number of specimens; but I now took to examining them more closely, and drawing them to scale.

The principal papers already published on this subject are 1. By Purkinje, in Latin (Vratislaviæ, 1830, 4to). His illustrations of the cells of the anther are very interesting; but, unfortunately, the pollens drawn have been treated with water, so as to lose most of their distinguishing marks. 2. By Mirbel, in French (Elémens de Physiologie végétale,' 1815). Accurate as far as they go, but on too small a scale.

3. By Fritzsche, in German (Berlin, 1832, 4to). Most beautifully drawn both in the simple form, under oil, and under the effects of weak sulphuric acid. He examined almost all the natural families.

4. By Hugo Mohl, in German (Berne, 1834, 4to), afterwards translated into French in the Annales des Sciences Naturelles,' 1835. He goes most thoroughly through all the families in the natural state of the pollens, or as they appear under olive oil or water; drawn to scale. I have followed in his footsteps; and now I wish to point out a few inaccuracies into which he has fallen, and to carry on the examination of the other plants.

6

5. Hassall reviewed Mohl in the Annals of Nat. Hist.' 1842, vols. viii., ix., showing some differences, and going through all the natural families in detail, with numerous plates drawn to scale.

B

6. Lindley, in his 'Introduction to Botany,' published figures of pollen after some of the above authors, and

7. A further republication of them in the 'Micrographical Dictionary.'

8. Since that time Hermann Schacht has published some beautiful drawings in Pringheim's 'Jahrbücher für wissenschaftliche Botanik,' Berlin, ii. p. 109, 1860. I agree very much with what he has published, except about the Pinus family, of which his descriptions differ from what I have observed, and from the drawings of Mohl, and Hassall, and Dr. Hooker.

9. Nägeli has also a paper in the same publication, vol. iii. Rosanoff has a paper in the same publication, iv. p. 34 (1865), principally on the pollen of the Acacia.

10. Luerssen, in vol. vii. p. 34 (1869) of the same publication, reviews a paper by

11. M. Pollender, Bonn, 1867–68.

12. A paper by A. W. Bennett in the 'Popular Science Review,' April, 1875; and lastly,

13. Some woodcuts by Worthington Smith in the 'Gardeners' Chronicle' in October 1876; republished in the 'Monthly Microscopical Journal,' January, 1877.

The pollen forms are often noticed by Dr. Hooker in his large works, and also by some foreign botanists in the Brazil flora.

It is remarkable, that while in some families the species vary greatly, in others they are almost always the same. Some may therefore be considered normal, others as having only a specific value. The different size of the grains is also marvellous-some very large, and visible to the naked eye, as in Stachytarpheta, the Malvaceæ, Cobæa, Pancratium, and some species of Iris, reaching as much as ro of an inch; while others, as many Boragineæ, Primulaceæ, Melastomaceæ, and Saxifrageæ, are as small as 1 to 2-6000ths of an inch. In colour also they are very various, but usually some shade of yellow or white; but I have observed blue, red, brown, or (rarely) green, which has been denied by Lindley.

Mr. Bennett would divide pollens into those which are windborne, or anemophilous, and those which are not so scattered, or entomophilous; but the grains are by no means all globular, as he asserts. The Gramineæ and Cyperaceae and Humulus are usually prismatic. The Populus has globular pollen; while the pollen of the Salix, in the same family, is somewhat cylindric,

as is that of Rheum (which is probably anemophilous) and of the dioecious Rumices, with three slits.

Plantago pollen, which appears to be also anemophilous, varies from a polyhedric to the simple form, while the wind-borne pollen of the Coniferæ is very different.

The pollen grains of Urtica are very small and elliptic, with three slits, withering into a prismatic form.

Those of the Grasses and Cyperaceae, and perhaps the Plantagineæ, are without the sticky nature of the outer coat, which obtains through all other pollen grains.

Some grains throw out a quantity of tubes, by which they are hung together, as in Richardia ægyptiaca, Strelitzia regina, and the Rhododendron, &c. ; but more usually they are quite distinct. I omit any account of the peculiar pollen of the Asclepiadeæ and Orchidea, which have not common loose grains of pollen, but joined together in a coherent mass. I also do not intend to make any remarks as to the structural nature and function of the fovilla and the pollen grains.

I have found but three or four instances in which the same anther produces two or three forms of pollen. Usually they are all similar, often with some small or imperfect grains; but in Mimulus luteus I have found the simple form, common in the Scrophulariaceæ, mixed with the coiled form, which H. Mohl pointed out in Mimulus moschatus (No. 138). I have not been able to find this coiled form myself in the latter; and Mr. W. Smith considers it an error of Mohl: but in this he is doubtless mistaken; for I have found it in the next species, M. luteus. I have also found many forms of pollen in Browallia elata (No. 146), in the same anther, but not in the coiled form; and Mohl mentions others also which I have not had opportunities of examining.

I find also two forms in Thalictrum flavum and T. glaucum (Nos. 430, 433), probably also in Ranunculus auricomus, as remarked by Hassall and myself in the Anigozanthos and some Rubi.

The simplest form is a bag with one slit, from which the fovilla escapes, then opening out into a hat-shape, as in many of the Liliacea and Amaryllideæ.

A similar form is seen in the Magnolia and Nuphar—a simple slit. It is curious that the simplest pollen should be found in one of the most ornate and largest of flowers.

This simple form is often found with ornate markings on the outside, in the Liliaceæ, for instance.

The next simplest form looks at first the same; but on examination it is found to contain three slits or three bands, the bands more or less broad. When floating in olive oil, so that all the sides of the grain are shown, the end is usually somewhat three-cornered, consisting of the ends of the three slits or bands. This is universal, as far as I have seen, in the Saxifrageæ and Crassulaceæ, and almost universal in the Rosacea: the Poterieæ (Nos. 286-7) and Spirea palmata (No. 285) are the only exceptions I have seen. Hassall mentions another form with four slits in Rubus. I at first found only the usual form with no variety but in size, in many of the common subspecies of R. fruticosus, the discolor or cæsius; but at last I have found in R. radula a few of the four-slit form, with the common form both larger and smaller and similarly a few in R. hystrix, and still fewer in R. dumetorum; but in all, the majority of the pollen was of the usual form. The Scrophulariaceae, with the exception of Mimulus (which has three forms), most of the Solanaceæ, and all that I have examined of the Melastomaceæ are similar, but are the smallest I have seen of this form. It is found in many other families, in which, however, there are also other forms; as in some Ranunculaceae and Sapindaceae, and in Spergularia rubra (No. 360), and Spergula alone of the Caryophyllaceae. The Primulaceae have usually three slits, and when seen end-up the pollen looks trigonous; but in some cases they have four or six slits, as remarked by Mr. A. W. Bennett in the 'Popular Science Review' of 1875. In P. cortusoides (No. 189) it is either simple or trigonous, as I have examined, and as described by Mohl. The last-named and Hassall both describe the pollen of the Gramineæ as subglobular, with one pore; but I have examined numerous specimens, and find it (with the exception of the Paniceæ, of which I have examined but three) always of one peculiar type-of a somewhat prismatic form, with rounded ends, larger at the top, and smaller at the bottom; while there are four sides with an oval or round opening in them, occasionally 3-sided, 5-sided, 6-sided (Nos. 1–4). Mirbel describes it in Phleum as I have observed it. Mr. A. W. Bennett describes this form in Briza, but says the Dactylis pollen is oval; but I think he cannot have floated it in oil, or he would have seen the same form as in Briza. I fancy that Mohl and Hassall chanced merely to see the upper side.

Arundinaria, Lagurus, and Pogonatherum have a 6-sided prism, as represented by Mr. W. Smith. The only Panicum I

have examined is P. plicatum (No. 5), in which it is quite different from that of any of the other Gramineæ, being globular with three pores; and there is something of the same kind in Streptium, a South-American genus of the same suborder. The Cyperaceae have two forms—one like a drop of water (in Cladium and Scirpus); in Carex it is somewhat prismatic, with a pore in each facet, which is the form in the Juncaces. That of Drimyopsis in that family is peculiar, like a tooth, a deformed crystal of the Carex type (No. 23).

To return to the more banded forms. The Papilionaceæ have generally three broad bands; when seen endwise the pollen looks trigonous. Erythrina and Indigofera have really a trigonous form, not cylindrical. Phaseolus has spherical pollen with 3 pores, which sometimes assumes a subtrigonal appearance. Some are

broader than others, but of a similar type, which obtains also more or less among the Casalpinieæ, with the exception of Bauhinia, in which it is of a very different type, large and spherical, and echinulate, or large and trigonous. The pollen grains of the Acacies are (with the exception of Mimosa pudica and a few other species) composed of nine to twelve or sixteen cells, in squares or pyramids, or a cubical arrangement. The Cruciferæ have also three bands, smooth or rough, usually containing much oil, which is diffused in the olive oil. In Schizopetalon it is rougher, very richly marked, with no distinct slits (No. 385). Mr. Bennett has remarked the globular form of Pringlia; but it is possible this form has been assumed when examined under water, or having been preserved in spirit. The Apocynaceæ, Jasmineæ, and Rutaceæ, in general have 3-banded grains, often with rich ornamentation on the outer coat, and containing much oil, which is diffused in olive oil, as is also the case with the very oily Tiliacea pollen. Adenandra fragrans, in Rutacea (No. 315) has two forms, one slightly curved, the other having an additional coat across the lower one: this is also found in the Umbelliferæ, which have also three slits, but are generally constricted in the centre. There are several forms of trigonous pollen: those of the Myrtaceae (as far as I have observed), the Proteaceae (except Banksia, which has two instead of three pores at each end), many of the Verbenaceæ, as Verbena (No. 105) and Lippia pollen; while Stachytarpheta pollen is of large size ( of an inch), with its three pores containing large craters, figured by Fritzsche and Hassall.

The Hydrophyllacea and most of the Cucurbitaceae-finally,