ley's Riverside Natural History, in six volumes, and his desire is satisfied.

In dealing with books of reference, there will often be found very important discrepancies of statement, different works giving different dates, for example, for the same event in history or biography.

Next to a bible and a dictionary of language, there is no book, perhaps, more common than a biographical dictionary. Our interest in our fellow-men is perennial; and we seek to know not only their characteristics, and the distinguishing events of their lives, but also the time of their birth into the world and their exit from it. This is a species of statistics upon which one naturally expects certainty, since no person eminent enough to be recorded at all is likely to have the epoch of his death, at least, unremarked. Yet the seeker after exact information in the biographical dictionaries will find, if he extends his quest among various authorities, that he is afloat on a sea of uncertainties. Not only can he not find out the date of decease of some famous navigators, like Sir John Franklin and La Perouse, who sailed into unexplored regions of the globe, and were never heard of more, but the men who died at home, in the midst of friends and families, are frequently recorded as deceased at dates so discrepant that no ingenuity can reconcile them.

In Haydn's Dictionary of Dates, Sir Henry Havelock was said to have died November 25th, 1857, while Maunder's Treasury of Biography gives November 21st, the London Almanac, November 27th, and the Life of Havelock, by his brother-in-law, November 24th. Here are four distinct dates of death given, by authorities apparently equally accredited, to a celebrated general, who died within forty years of our own time. Of the death of the notorious Robespierre, guillotined in 1794, we find in Chalmers'

Biographical Dictionary that he died July 10th, in Rees's Cyclopaedia, July 28th, and in Alison's History of Europe, July 29th. Doubtless it is some comfort to reflect, in view of his many crimes, that the bloody tyrant of the Jacobins is really dead, irrespective of the date, about which biographers may dispute. Of the English mechanician Joseph Bramah, inventor of the Bramah lock, we learn from the English Cyclopaedia, that he died in 1814, and from Rose's Biographical Dictionary, that he died in 1815.

Now, although a large share of the errors and discrepancies that abound in biographical dictionaries and other books of reference may be accounted for by misprints, others by reckoning old style instead of new, and many more by carelessness of writers and transcribers, it is plain that all the variations cannot be thus accounted for. Nothing is more common in printing offices than to find a figure 6 inverted serving as a 9, a 5 for a 3, or a 3 for an 8, while 8, 9, and 0, are frequently interchanged. In such cases, a keen-eyed proof-reader may not always be present to prevent the falsification of history; and it is a fact, not sufficiently recognized, that to the untiring vigilance, intelligence, and hard, conscientious labors of proof-readers, the world owes a deeper debt of gratitude than it does to many a famous maker of books. It is easy enough to make books, Heaven knows, but to make them correct, "Hic labor, hoc opus est."

A high authority in encyclopaedical lore tells us that the best accredited authorities are at odds with regard to the birth or death of individuals in the enormous ratio of from twenty to twenty-five per cent. of the whole number in the biographical dictionaries. The Portuguese poet Camoens is said by some authorities to have been born in 1517, and by others in 1525; a discrepancy of eight years.

Chateaubriand is declared by the English Cyclopaedia to have been born September 4th, 1768; September 14th, 1768, by the Nouvelle Biographie générale of Dr. Hoefer; and September 4th, 1769, by the Conversations-Lexicon. Of course it is clear that all these authorities cannot be right; but which of the three is so, is matter of extreme doubt, leaving the student of facts perplexed and uncertain at the very point where certainty is not only most important, but most confidently expected.

Of another kind are the errors that sometimes creep into works of reference of high credit, by accepting too confidently statements publicly made. In one edition of the Dictionary of Congress a certain honorable member from Pennsylvania, in uncommonly robust health, was astonished to find himself recorded as having died of the National Hotel disease, contracted at Washington in 1856. In this case, the editor of the work was a victim of too much confidence in the newspapers. In the Congressional Directory, where brief biographies of Congressmen are given, one distinguished member was printed as having been elected to Congress at a time which, taken in connection with his birth-date in the same paragraph, made him precisely one year old when he took his seat in Congress.

Even in reporting the contents of public and private libraries, exaggeration holds sway. The library of George the Fourth, inherited by that graceless ignoramus from a book-collecting father, and presented to the British nation with ostentatious liberality only after he had failed to sell it to Russia, was said in the publications of those times to contain about 120,000 volumes. But an actual enumeration when the books were lodged in the King's library at the British Museum, where they have ever since remained, showed that there were only 65.230 volumes, being but little more than half the reported number. Many libra

ries, public and private, are equally over-estimated. It is so much easier to guess than to count, and the stern test of arithmetic is too seldom applied, notwithstanding the fact that 100,000 volumes can easily be counted in a day by a single person, and so on in the same proportion. Here, as in the statistics of population, the same proverb holds good, that the unknown is always the magnificent, and on the surface of the globe we inhabit, the unexplored country is always the most marvellous, since the world began.

These discrepancies in authorities, and exaggerations of writers, are not referred to for the purpose of casting doubt upon all published history, but only to point out that we cannot trust implicitly to what we find in books. Bearing in mind always, that accuracy is perhaps the rarest of human qualities, we should hold our judgment in reserve upon controverted statements, trusting no writer implicitly, unless sustained by original authorities. When asked to recommend the best book upon any subject, do not too confidently assert the merits of the one you may think the best, but say simply that it is well accredited, or very popular. It is not always safe to recommend books, and the librarian does well to speak with proper reservations as to most of them, and to recommend only what are well known to him to be good, by his own intimate acquaintance with them, or, which is the surest test of all, by the verdicts of critical reviews, or by the constant reprinting of them in many successive years.

It was the well-nigh unanimous report at a Conference of American librarians, upon the subject of "aids to readers", that "nothing can take the place of an intelligent and obliging assistant at the desk." This was after a thorough canvass of the relative merits of the various reference books and helps to readers in book form. Not only the casual

reader, and the reader with a purpose may be constantly aided by the librarian's knowledge, and larger experience in the art of finding things, but teachers in the schools, clergymen preparing discourses, and every one seeking to know anything, should find the librarian a living catalogue. There is nothing so effective in the world as individual effort.