been indicated above, and has proposed the following method, the application of which he has exemplified by reference to several ancient English yews.

The girth is to be taken under the swelling which supports the first great branches, and where the trunk is free alike from bumps which exaggerate, and from twigs and spray which prevent, accurate measurement. At this level the trunk is to be perforated about 3 inches with a trephining instrument at the two opposite ends of a diameter, and also at an intermediate point; the rings are counted on a longitudinal section of each cylindrical piece thus extracted; and the mean of the three is taken for the number of rings per inch of radius removed. This would be the sole factor by De Candolle's formula, allowance being made for the greater width of the first 40 to 70 rings. But Bowman provides several intermediate factors from trees of several known ages, ascertained either by the date of plantation, or by the total rings on a complete trunk section, or by the method now under consideration.

For example, let the question be the age of an oak 16 feet in girth at 5 feet from the ground; and let it be understood that by observation various oaks, grown in similar circumstances with itself, are known to be 60 years old if 5 feet in girth, if 9 feet 120 years, and if 12 feet 170. It remains to determine the number of rings on the prolongation of the radius corresponding with the 48 inches which the circumference of the tree of 16 feet in girth exceeds that of 12 feet,—that is, a radius of 7.64 inches. But the mean of three borings has given, say, 60 rings for the outermost 3 inches, and therefore 153 for the whole radius. Consequently the age of the tree is 323 years.

This method will obviously give in general better results than the simpler method of De Candolle. But it too is open to objections. In the first place it proceeds on the assumption that a series of younger trees are in their growth a measure for an older tree at the several periods of its life corresponding with their several ages; whereas, even if we had a mean of several observations at each of these ages, that mean might be far removed from the actual growth of another tree of the same species, circumstanced more favour

ably for vigour or the reverse. This, however, Mr Bowman very well knew; and indeed the objection manifestly holds in regard to every possible method of estimating age except the actual counting of the whole woody rings of a complete section of the tree under examination. Another objection which Mr Bowman seemed to have overlooked, is that, as the ratio of rings per inch of radius is taken from the outer ones only to calculate many more within, though there is a progressive decrement of their width in all old trees,-the ratio, being taken from the most crowded part of a given portion of radius, must necessarily give as the result a greater age than the reality. The error thus arising will of course be the less, the greater the number of factors which have been furnished for the calculation by treating the younger trees, as well as the outer radius of the oldest, by Bowman's method. Thus in the case of the Douglas pine mentioned above (p. 227), the section of which presents a very regular decrease in the radial width of its rings, the actual number is 303; but calculated from the ratio of radial width of the last ten years of each successive fifty after the first, the result is 330. It may be thought that this amounts to a fair approximation to the truth in such a question. A third objection, which may often prove fatal to Mr Bowman's method, is that, however suitable for a traveller's observations in a wild country, it is not always that, in a civilized country, the owner of a fine tree will consent to deep holes being bored into its trunk with a trephining instrument.

*

This objection cannot apply to the method for ascertaining the radius for a given number of rings by relative measurement of the girth at a known interval of years. It is true

Radial width of 50 Radial width of last Rings calculated from

Age.

19.14 inches.

330.0 years.

that, in the case of young and even middle-aged trees, an inconveniently long interval between the measurements, probably ten years, may be necessary to exclude the chance of error from the influence of special favourable or unfavourable seasons upon tree-growth. But for aged trees a much shorter interval will suffice, because the width of their narrow outer rings appears to vary little, one year with another, if the measurement be kept clear equally of the conoidal base or buttresses of the trunk, and of the swelling spring of the lowest great branches. Moreover, as the whole girth is used for deducing the radius, this method is likewise exempt from the risk of error in employing for that purpose even the mean of three random borings.

Another peculiar application of the measurement of the successive layers of wood is to ascertain the relative nature of seasons long gone by from observation of the relative width of the layers. This has been a favourite speculation with many; but it is more specious, I fear, theoretically than sound in practice. Theoretically it seems a true proposition. But it assumes that the circumstances in which the tree is placed have been for a long term of years the same in all respects except as to season; and unfortunately the case is for the most part directly the reverse. Differences in the amount of annual increment cannot fail to arise when in the progress of time the roots happen to reach a better or worse soil, or when in spreading they come into competition with the roots of neighbouring trees, or are supplied by accident or designedly with better or worse terrestrial nourishment in some other way, or when the foliage becomes cramped on the one hand, or expansive on the other, by the growth or removal of its neighbours, or their neighbouring branches. To these variations, much more than to seasons, must be assigned the frequent observa tion, that the relative width of the rings varies in one or more of them in succession at different points of the circle, or even in the whole circumference. Accordingly I have several times tried in vain to trace the effect of known seasons on a section of a tree. Let those who believe this to be practicable endeavour to trace the wonderful season of 1826, when warmth and sunshine, commencing with March, ended only with September, and when the summer was

continuously such as to change in some respects the habits of the people; and should they succeed, the hypothesis may be worthy of further investigation. The best chance of success would probably be afforded by choosing a tree in the middle of a deep wood, or standing singly in an open field, and which has been long treated uniformly and judiciously.

XXII. On the Flora of Mull. By Mr George Ross.

(Read 9th May 1878.)

As Mull has been less visited by botanists than the other large islands in the west of Scotland, and as it was stated in the Report of the Botanical Record Club, 1873, that no account of its flora existed, I sent to the Club in 1876 a list of the flowering plants, ferns, and their allies, observed during my residence. This was printed in the Report of the Club for that year, but with only a few of the stations. The following list contains a considerable number of additions, but is still far from complete, as I have not yet been able to visit a large portion of the island, including the highest hills; this may explain the absence of many alpine species found on hills of the same height elsewhere in Scotland.

The highest hills are Ben More, 3172 feet; Dun-da-gu, 2505; Creichbeinn, 2344; Ben Buy, 2332; Ben Greig, 1941; and a number of others are between 1000 and 2000 feet. It is all" up hill and down dale,"-no extent of level ground. Scattered among the hills are numerous lochs, the largest of which, Loch Frisa, is 5 miles long, Loch Baa about 3; but the greater number are of small size, though often very deep, and contain excellent water swarming with trout.

The geological formation is invariably trap, except the termination of the Ross, which is granite, and patches of Tertiary strata at Ardtun and Griban. Ben More and the hills near it gave origin to all the traps of Mull and the islands around it, including Staffa and Iona, besides large tracts of the adjacent mainland. In those traps denudation has been, and is still, going on at an enormous rate; so fast are they crumbling away, that the debris and stones carried down to the hollows or the sea are so frequently covered over by a

fresh layer, that they have not time to form sand or rounded pebbles by the action of the waves; consequently, the sea bottom is mostly formed of angular stones. This waste from the rocks also constitutes all the soil except the peat, which is abundant all over the island.

The shores are in general rough and often precipitous, and in many places difficult to approach. The west coast is, however, cut into by a number of lochs, and their shores are in general easy of access, the two largest having a road near the shore nearly all round. On the same side at Griban and Ardmeadonach are the highest sea-cliffs on the island. On the shore at Calgary Bay and several parts of Loch Cuan is a snow-white sand (if the term be applicable to a substance entirely composed of broken shells), on which are a number of plants found nowhere else on the island. have seen no silicious sand except on the granite of RossMull.

The great bulk of the island is in a state of nature, and there are some extensive woods. Although the geological formation is chiefly trap, so many varieties of this class of rocks occur that there is a great difference in the character of some localities from that of others; yet, the soil being invariably of the same nature, no great difference in the vegetation. except from altitude is to be expected.

As the list gives a fair idea of the frequency of species, I shall only mention a few I had considered rare, but which are quite the reverse in Mull. Trollius europaeus is plentiful all over the north, and seemingly so in the south; I found it still in flower at Beach, Loch Scridain, on the 18th July. Drosera anglica is equally common; in some places it seems to have nearly starved out D. rotundifolia. Pinguicula lusitanica is by most of the burn sides and springs; Pyrola media is frequent-I have not seen it elsewhere on the west of Scotland; and the rare Scutellaria minor is probably not uncommon-I found it both on Loch Keil and Loch Scridain in abundance; Schænus nigricans and Rhyncospora alba are both common, going a long way up the hills; Calamagrostis Epigejos is frequent; and many others might be added that range from end to end of the island, which are not usually so common in other parts of the country.

I have seen white varieties of Epilobium montanum,