When this book was begun, some years ago, I made a formal plan, according to which it was to have been one long Essay or Treatise, divided into sections and chapters, and presenting that apparently perfect ordonnance which gives such an imposing air to a work of art. I say “apparently perfect ordonnance,” because in such cases the perfection of the arrangement is often only apparent, and the work is like those formal pseudo-classical buildings that seem, with their regular columns, spaces, and windows, the very highest examples of method; but you find on entering that the internal distribution of space is defective and inconvenient, that one room has a window in a corner and another half a window, that one is needlessly large for its employment and another far too small. In literature the ostentation of order may compel an author to extreme condensation in one part of his book and to excessive amplification in another, since, in reality, the parts of his subject do not fall more naturally into equal divisions than words beginning with different letters in the dictionary. I therefore soon abandoned external rigidity of order, and made my divisions more elastic; but I went still further after some experiments, and abandoned the idea of a Treatise. This was not done without some regret, as I know that a Treatise has a better chance of permanence than a collection of Essays; but, in this case, I met with an invisible obstacle that threatened to prevent good literary execution. After making some progress I felt that the work was not very readable, and that the writing of it was not a satisfactory occupation. Whenever this happens there is sure to be an error of method somewhere. What the error was in this case I did not discover for a long time, but at last I suddenly perceived it. A formal Treatise, to be satisfactory, can only be written about ascertained or ascertainable laws; and human intercourse as it is carried on between individuals, though it looks so accessible to every observer, is in reality a subject of infinite mystery and obscurity, about which hardly anything is known, about which certainly nothing is known absolutely and completely. I found that every attempt to ascertain and proclaim a law only ended, when the supposed law was brought face to face with nature, by discovering so many exceptions that the best practical rules were suspension of judgment and a reliance upon nothing but special observation in each particular case. I found that in real human intercourse the theoretically improbable, or even the theoretically impossible, was constantly happening. I remember a case in real life which illustrates this very forcibly. A certain English lady, influenced by the received ideas about human intercourse which define the conditions of it in a hard and sharp manner, was strongly convinced that it would be impossible for her to have friendly relations with another lady whom she had never seen, but was likely to see frequently. All her reasons would be considered excellent reasons by those who believe in maxims and rules. It was plain that there could be nothing in common. The other lady was neither of the same country, nor of the same religious and political parties, nor exactly of the same class, nor of the same generation. These facts were known, and the inference deduced from them was that intercourse would be impossible. After some time the English lady began to perceive that the case did not bear out the supposed rules; she discovered that the younger lady might be an acceptable friend. At last the full strange truth became apparent,—that she was singularly well adapted, better adapted than any other human being, to take a filial relation to the elder, especially in times of sickness, when her presence was a wonderful support. Then the warmest affection sprang up between the two, lasting till separation by death and still cherished by the survivor. What becomes of rules and maxims and wise old saws in the face of nature and reality? What can we do better than to observe nature with an open, unprejudiced mind, and gather some of the results of observation?