Correspondence — Mr. Hugh Miller. 
41 
we occupy not unlike grounds. Mr. Bonney admits some tarns, like 
Grasmeve, as glacial ; admits, hesitatingly, in part if not wholly, 
some lakelets, (do tbese include Grasmere’s neighbours ?) ; admits 
that once a basin is formed, a glacier works in it “under very 
favourable couditions” (Letter to Mr. Fisher, p. 377), thus granting 
to tbe process increasingly favourable conditions ; but demurs to tbe 
statement “ that though competent to deepen a lake-basin, a glacier 
could originate it.” It would tbus seem to be with Mr. Bonney 
a question, not of ability, but of time.” Did the glacial period last 
long enough to eDlarge, under “very favourable conditions,” a tarn 
it was able to originate under less favourable conditions. 
If I understand Mr. Bonney correctly, we are at one both in our 
desire to bring eacb case to the test of observation, and in our ap- 
preciation of tbe increase of tbeoretical probability as the series 
advances from tarn to lake. But if the utmost that even bis careful 
observations can do for bim is to render tbe glacial theory probable 
or improbable (p. 376), tben surely tbese theoretical probabilities 
are worthy of greater weight than be gives tliem. His illustrations of 
blown sand eroding (must we say tarns ?), and Homeric youtbs spread- 
ing erratics, seem to me scarcely relevant to tbe state of tbe question. 
Tbe latter paragraphs of Mr. Bonney’s letter call for no remark 
from me, as tbey involve — at tbis stage — a knowledge of the Alpine 
lakes that I do not possess. I may assure Mr. Bonney, however, 
that though I bave ventured to remark on bis theories, I do not 
question bis facts. 
Tbe letter of my friend Mr. Judd I must attempt — with niucb 
diffidence in my own powers — to answer, for it involves destruction 
to my position. First, let me say a word of explanation. In sup- 
posing me prepared “ to admit tbe overwhelming probabilities ” of 
tbe subsidence theory in regard to all the larger lakes, Mr. Judd 
misunderstands me. A priori probabilities in relation to lakes both 
large and small, I believe must be conceded to both tbeories. But in 
such questions, overwhelming probability can be allowed only to 
overwhelming proof. 
In the second place, as regards the halting-place that Mr. Judd 
finds between tarns and lakes. If, as I argued, a glacier is a tool 
that greatly grows in calibre and efficiency as a tarn-hollow en- 
larges, — that scrapes harder and scores deeper, tben to concede 
tarns to tbe feebler tool and deny them enlargement by tbe mure 
powerful, notbing being pointed to as intervening to stop the action, 
is what may very properly be characterized as not logical and not 
reasonable. Nevertheless, as for want of standing room and a 
fulcrum, Arcbimedes found bis theoretically infinitely powerful tool 
limited by “reasonable proportions” — the limit of all terrestrial 
tools, so are “ reasonable proportions ” the limits also of glaciers and 
their work. Wbile caution tben compelled me to remember, and to 
indicate in my paper, that there are limits to tbe enlargement of 
tarns by glaciers, that fact — even in Mr. Judd's able hands — leaves 
tbe tarn and lake question precisely on its former basis. 
My friend suggests to me, however, au aualogy which may belp 
