140 



♦ KNOV^^LEDGE ♦ 



[Aug. 31, 1883. 



But it is clear that the triangles e « E' andya F' are also 

 similarly proportioned to B « 6 of Fig. 8. For a e, like 

 a b, is truly horizontal, and a E' like a B is directed to the 

 sea-horizon. It matters not in what direction we look 



the mirror experiment of page 101, that, with a one-foot 

 mirror and two yards distance would command an arc of 

 less than five degrees of sea-horizon, and it would be pre- 

 posterous to expect the slightest trace of curvature along 

 such an arc. 



(To he continued.) 



seawards from a fixed point above the sea ; the sea-horizon 

 has always from such a point the same depression. The 

 only difference is that a e and a f are longer than a h, 

 while e E' and J'Y are longer than h B in the same 

 proportion. 



Whereas then B i is about 1 ^ in. in length, e E' and JY 

 exceed Ig in. in the same degree that «« or a/ exceeds 

 ah ; which is (appreciably) the same degree in which a E 

 or a F exceeds a B, that is as the diagonal exceeds the side 

 of a square (for a B and B E are equal and at right angles 

 to each other). Thus since the diagonal of a square is 

 about 1'414 when the side is 1, we have 

 «E'=/F'= IJ in. X 1-4U 

 and e E =./"F = 1^ in. (each being equal to B 6) ; 

 .". E E' = F F' = \\ in. x 0-tl4 = -.5.5 in. approximately 



In reality, taking refraction into account the angles of 

 depression are all reduced by about one-fifth, leaving E E' 

 and F F' optically equal to only -W in., or ^ths of an inch. 

 The actual curvature of E' B F' would be fairly shown if 

 in such a diagram as Fig. 14, E B and B F were each 2-5 ft. 

 long, E E', B H, and F F' each Jths of an inch long ; and 

 if then the curve E B F were swept out in the narrow 

 rectangle E F', whose length would he more O'ln thirteen 

 hundred times its hreadth. 



So that even from so great a height as 2(. ft, a ridge 

 roof so long as 50 ft., seen from a distance of 25 ft., com- 

 manding therefore a range of a full right angle along the 

 roof, and brought at the two extremities to exact apparent 

 coincidence with the sea-horizon would be less than half 

 an inch below the sea-horizon at its middle point, even if 

 use were made of such a point as a (Fig. 13) to guide the 

 eye. 



But all the observations ever made in this way, as by 

 Parallax in 1864, by Mr. Hardy (see pase 77), and by a 

 few perplexed students of this matter, have been made 

 from lower levels than 200 ft , on ridges commanding a 

 much smaller angle than a right angle, and without any 

 such a point as a to guide the eye, though such a point is 

 absolutely essential to exactness of observation. As for 



PUNCTUATIOX AND PRINTERS. 



I "WONDER what the Brighton Herald takes to be "the 

 rights and privileges of the humble comma." Appa- 

 rently the right of distraction ; for everything that unduly 

 stops the continuity of thought in reading is distraction — 

 parentheses, whether written so or not, unnecessary or too 

 strong stops, all sorts of carts before horses, exceptions 

 and conditions stated at the beginning instead of the end 

 of a sentence, verbs before their nominative cases or after 

 their accusatives, except for special reasons, anything that 

 obliges you to wait till the end in order to realise what 

 the sentence is about. These things, and not the mere 

 length of sentences are the impediments to easy reading. 



I forget whether you noticed before the printers' passion 

 for dUMing in a comma between every two adjectives not 

 alread}' separated or joined by an and, without the least 

 regard to its effect. For instance somebodv might write 

 that I am a "good tall man," meaning only that I am 

 above the average height : the average printer would forth- 

 with exalt me into a "good, tall man"; and thereby would 

 raise the indignation of every architect in England into a 

 still hotter flame than it is in just now I see, at my having 

 dared to rebuild part of a cathedral without paying them 

 some £1,500 of black-mail for doing worse than nothing, 

 or something like the Law Courts. 



But I want to go beyond commas, and to give even the 

 Devil his due. Some men deliberately leave their punctua- 

 tion to that inky Diabolus, and it is well for their readers 

 and their own credit that they do. Within the last month 

 or so I have had letters from two men of excellent educa- 

 tion, lioth High Wranglers and men of science — one old 

 and the other young — whose only stops were commas and 

 dashes, dashed in pretty much ad libitum. Such writing 

 would justify quite natural language from a printer's devil ; 

 and if he is to be condemned himself for doing his best — 

 and rather overdoing it — he may fairly call the world un- 

 grateful. But, on the other hand, he must see in five 

 minutes whether he is printing for a punctual punctuator 

 like you or me, who, at any rate, mean to put in all our 

 stops ; and when he sees that he ought to leave us alone 

 to bear our own iniquities if we let our proofs go "un- 

 peppered " more than he thinks right. We probably know 

 better than he does. 



The fact is that any one who attempts either to 

 punctuate or to write English entirely by rules is sure to 

 disregard both sense and common sense sometimes. But 

 one rule may be safely followed, viz., the advice of some 

 judicious bishop to his clergy on the length of sermons, to 

 " err on the side of leniency ;" and the other still older 

 rule — Quid duhitas nefeceris, if you doubt about a comma 

 leave it out. 



The Greeks were wiser than we are in not encumbering 

 themselves with both colon and semicolon. Xobody can 

 say with certainty i'lnot TrXiuv rjftiav -tiiroc, how much less a 

 semicolon is than " the entire animal," from which it 

 differs only by a tail. What is really wanted is a kind of 

 minor comma, or both that and the ieniioolon reducing a 

 little in power of obstruction. But that reform is as 

 hopeless as an Easter by the sun instead of the (Act of 



