This very modest effort is not intended to advance the study of Canonical chronology, but rather to offer some comfort to those who wish to take their Canon 'straight' (that is to say, without an excess of watering-down commentary). It is my hope that it achieves this by showing that what Watson provides us by way of dating is both reasonable and respectable.

	Some Chronological Crankiness A chronology for the calendrically challenged
I act only from a sense of public duty		On refering to my notes, I see that ...

THE CHRONOLOGICAL PROBLEMS OF THE CANON are such that no two chronologists seem ever able to agree and the impression is given that the Chronicler himself has provided sufficiently contradictory data so that this state will persist for some time ('forever' is not ruled out). I believe this impression to be false, and that the Chronicler has provided consistent, if not always adequate, data. It may be that what we have here is a 'trees and forest' problem. In the effort to pin down a date certain for each case in the Canon, chronologists tend to focus in on the subtlest of clues and hints (so far as I know none has yet investigated the sink-rate of parsley in butter at varying temperatures so as to obtain a clue as to the time of year), perhaps overlooking and/or underweighing the broader aspects of the Canonical time-line.
    What I propose to undertake is what might be termed a topological, rather than metrical, approach to the question, seeking primarily to establish the order of the cases. Dates (in the sense of day of the month and day of the week) will not come into it at all. No doubt this will leave some lacunae, but this outcome can be no worse than what now appertains. (You may skip the argument and view a table of results directly. A somewhat formal setting out of the criteria is also available.)
    In determining the order of the cases without reference to dates, one looks to internal evidence which can establish the precedence of one case over another. A simple example can make this clear. In A Scandal in Bohemia (hereinafter SCAN), Watson remarks that he "had seen little of Holmes since the singular chain of events which I have already narrated in a bold fashion under the heading of The Sign of Four." He goes immediately on to speak of the "home-centered interests" of the newly-married man, making it very clear that it is his marriage to Mary Morstan to which he refers. We thus obtain one 'topological' fact for our ordering: SIGN indubitably precedes SCAN, and no matter what the eventual place of either in the final ordering, SIGN must always precede SCAN. (And with very few cases intervening.)
    Such facts as these will determine the details of our chronology; however, we shall begin this effort by using external clues as our point of departure. The history of the publication of Watson's accounts reveals one major division: the long period from December of 1893 to August of 1901 when no cases were published. This immediately divides the cases into two groups of which we can say of the former that all cases published or mentioned must come before the temporarily tragic termination of Holmes' activities in May of 1891 at the Reichenbach Falls. For convenience we shall refer to these two major time divisions as 'B.R.' and 'A.R.', in honor of the dividing event.
    Of the 26 cases published B.R., 24 appeared in the relatively short span of time from July of 1891 until the aforementioned December (of 1893). Readers will not fail to note the close proximity in time of Holmes' supposed demise and the appearance of the first of the shorter accounts, suggesting that Watson may have intended these to be some sort of ave et vale for his friend.
    There is one event which can serve to divide these 24 B.R. cases: Watson's marriage to Mary Morstan. (Two of these cases, as we all know, precede in time the meeting of Holmes and Watson). In the last of these cases (FINA) he links 'my marriage' to "when I started in private practive," making it very clear that it is the Morstan marriage he has in mind. Thus, I believe we may reasonably infer that throughout the short three year period from Watson's 'plunge' (into matrimony) to Holmes' (into the falls), Watson was living apart from Holmes with the wife he wooed in SIGN. Consequently, if in the 24 short cases we find Watson living with Holmes, we may, with some confidence, place these cases in the ante-Morstan-marriage category. I would go somewhat further, however, and would maintain that this marriage is clearly his first, for if the wooing reminiscences of SIGN are at all authentic, Watson was something of a novice in these matters, tri-continental claims notwithstanding, so that ante-Morstan-marriage becomes simply ante-marriage.
    Applying this very minimalist criterion to the B.R. cases, I find I can construct the following list (preserving for the moment the order of publication for unresolved cases). I have supplied tentative years for some of the cases (wherever possible those which Watson has given); some are more questionable than others.

       GLOR (1874)
       MUSG (1878)
       ----- the meeting
       STUD (1881)
       SPEC, BERY, SILV, CARD, YELL, REIG, RESI, GREE,
       SIGN (1887)
       NOBL ('few weeks before marriage')
       ----- the marriage
       SCAN (1888 -- 'had not seen since marriage')
       IDEN (1888 -- 'had not seen for some weeks' -- after SCAN?)
       REDH, BOSC, FIVE, TWIS, BLUE, ENGR, STOC, CROO, NAVA, COPP*
       FINA (1891)

*COPP opens with a breakfast-table scene that seems to indicate that Watson is living with Holmes, and an earlier version of this chronology rather injudiciously placed it ante-marriage. But closer reading (well, reading) indicates that it must be post-marriage, for several of the cases of the Adventures are mentioned by Holmes, something he could hardly have done prior to Watson's marriage.

Added 3/05: We should note that, in terms of the initial criteria for this chronology, the cases of The Gloria Scott, The Musgrave Ritual, and A Study in Scarlet are unambiguiously ordered as the first three cases of the Canon, so that our effort below to assign dates is for convenience only, and to further a hypothesis I have about his early years.
    Holmes' age is critical to the first cases and our first step must be to settle on it. In His Last Bow Holmes -- disguised only by some chin-whiskers and an atrocious pseudo-American accent -- is characterized as "a man of sixty," with no qualifiying words such as 'looked to be' or 'had the appearance of.' Watson -- who is, of course, the author of this Epilogue -- will have known Holmes' age and has, I feel, simply given it to us. Obviously one cannot be certain of this, but I feel comfortable with it. If we accept 1854 as the year of Holmes' birth (the "traditional" January 6th is clearly a nominal choice for his specific birthday), we would conclude that he would likely have gone up to University when he would have been 18 or 19, which is to say in 1872 or 1873. He remarks on his "two years at University," indicating that he took no degree, as three years was the minimum residence required. This is mildly confirmed by his continuing to study at Bart's in London where Watson meets him in 1881 when he is 27.
    We know that during his years in London when he was residing in Montague Street he possesed an abundance of leisure time -- an over-abundance if his remark in The Musgrave Ritual is not an exaggeration. But by the time he meets Watson he seems to have developed a steady, if small, practice and to have become well-enough known to the police that they will call him in and he knows them well enough to know of the rivalry between Lestrade and Gregson.
    Holmes speaks of Victor Trevor as the only friend he made at University, and the events of The Gloria Scott take place during "the long vacation" when Trevor invites him home. But when Holmes leaves he goes up to his "London rooms," and we infer that the visit occured at the end of Holmes' two year's attendance. This would place Gloria in either 1874 or 1875. As there is no more data from which to build bricks, if we wish to assign a specific year we must chose arbitrarily, and if we opt for 1874 (because 18 seems to us a more likely age to matriculate), we need to keep in mind that it is symbolic only.
    The visit from Reginald Musgrave occurs four years after taking leave of the University, so if we chose 1874, 1878 follows by simple arithmetic I find this satisfactory, as it leaves something like two and a half years for Holmes, spurred by the visit from Musgrave (and perhaps with some assistance from the possibly-embarassed M.P. -- see here), to develop the "considerable" practice he claims (a slight exaggeration not untypical of the self-employed, one feels), and leaves four years of considerable leisure to fill with studies (Musgrave is only his third client, recall). It is during the four-year pre-Musgrave period that I postulate that he wrote some -- if not indeed many -- of his monographs for in A Study in Scarlet we find him telling Watson "I have made a special study of cigar ashes -- in fact I have written a monograph upon the subject." I rather think that for Holmes writting a monograph upon a subject followed quite naturally upon his making a (special) study of the topic (or science).
    For these reasons I have altered the dates for the first two cases to 1874 and 1878, respectively.

Inasmuch as the dating of SIGN has been called the 'lynch-pin' of Canonical chronology, I should probably indicate why I have accepted 1887 as its year (in this I follow the argument John Hall makes in "I remember the date very well"). Mary Morstan's father disappeared in 1878 "nearly ten years ago" (Emphasis added) Since 1882 Miss Morstan has received one pearl per year, now has six, and she naturally characterizes the time over which they have arrived as "about six years". But the pearls arrive in May; it is now September, and six pearls bring one up to 1887. (Just to be explicit, the pearls arrive in '82, '83, '84, '85, '86, and '87; hence the current year is 1887.) It is, of course, possible that 'it is now July', for Watson has offered both months. Which one chooses does not affect this argument, however.

The next level of clues we may now exploit in the Canon concern either a) references to the marriage (e.g., 'early in my marriage'), b) a specific year (which we shall use only to establish before/after relations), or c) the occasional reference to how long Watson had known Holmes. Taking the second un-ordered group first, many are 'dated' as 'shortly after my marriage' or 'the summer after my marriage', and this permits us to divide this group in two --- those shortly after the marriage, and those sometime later. This gives us the groupings:

       STOC, CROO, NAVA, ENGR
       REDH, BOSC, FIVE, TWIS, BLUE, COPP

The first three are definitely asserted to be in the summer after Watson's marriage, STOC in June, NAVA in July, and CROO in "the summer". STOC records Watson's re-entry into medical practice, and notes that he had seen litle of Holmes for about three months. This would correspond to the interval between SCAN (when he had presumably just taken over Farquar's practice) and the June date of STOC. This suggests -- rather strongly -- that the "few weeks" that he had not seen Holmes for in IDEN came after the three summer cases, and indeed we find Mary Sutherland enwrapped in "a heavy fur boa," surely indicative of the cooler weather of fall. We thus place IDEN after NAVA and CROO, which are definitely in the summer.
    ENGR is "the summer not long after" Watson's marriage. This argues for a longer interval than the following summer, and so we will provisionally place it in 1889. As to the other cases, REDH is said to have occurred in the autumn of 'last year', and a date of 1890 is provided. As it was published in 1891, this would seem to be a fairly good date. BOSC speaks of June 3rd being 'Monday last,' which would place it in 1889. TWIS and ENGR are explicitly dated to 1889. FIVE is remarked as occurring in 1887, but Watson is clearly married, which makes the date improbable as it would then be before our provisional date for the marriage; a misreading of a '9' for a '7' could account for the discrepancy, placing FIVE in 1889 (and, incidentally, making it possible for Irene Adler to be the woman who defeated Sherlock Holmes). BLUE has no time reference save its occurrence on the second day after Christmas, but does refer to several earlier cases, most notably TWIS, so it is perhaps the Christmas of 1889. COPP mentions TWIS, and is additionally noted as taking place on "a cold morning of early spring". As Holmes was occupied by Moriarty in the spring of 1891, it would seem that COPP takes place in the spring of 1890. If we accept this we can refine our groupings to:

       STOC, CROO, NAVA, IDEN (1888)
       BOSC, FIVE, TWIS, ENGR, BLUE (1889)
       COPP, REDH (1890)

Turning to the eight cases which seem to have occurred between STUD and SIGN, we find only two with years specified: SPEC, placed in April of 1883, and REIG, consigned to the spring (actually, April) of 1887. RESI is noted as coming at the end of the first year of life at Baker Street. If the year were 1882,its October month would make it over a year distant from when Watson moved to 221, his reference to 'year' must be to calendar year, so that it is the end of 1881 that he means. CARD is given as occurring during a hot August, supposedly no hardship to Watson, as his service in India had trained him to endure heat. This argues for a time not too long after his return from there, as his service in India was rather brief, and we would expect that acclimation to pass rather quickly. GREE has no date, but Watson's surprise at discovering that Holmes had a brother, and the fact that he speaks of their 'long acquaintance' suggests a time near the end of Watson's single days. BERY, SILV, and YELL contain no helpful references to time, and must remain unresolved. If we try to subdivide these cases, we get something like the following: (I shall now begin to add details to the year specification, so that the reasons for the specific order become clear(er).)

       RESI (1881 -- October)
       CARD (1882 -- not long back from India)
       SPEC (1883 -- April)
       BERY, SILV, YELL
       REIG (1887 -- April)
       GREE (1887 -- summer)

I have placed GREE last, and so in 1887, mostly because SIGN begins with Holmes' famous, if un-knowing, deductions about Watson's brother, and it seems possible that the discovery that Holmes had a brother may have prompted Watson to issue his injudicious challenge using his brother's watch which opens SIGN.

If we now assemble the pieces, we have for the chronology of the B.R. cases:

       GLOR (1874 -- summer)
       MUSG (1878 -- July)
       ---- the meeting
       STUD (1881 -- March)
       RESI (1881 -- October)
       CARD (1882 -- not long back from India)
       SPEC (1883 -- April)
       BERY, SILV, YELL
       REIG (1887 -- April)
       GREE (1887 -- summer)
       SIGN (1887 -- September)
       NOBL ('few weeks before marriage')
       ---- the marriage
       SCAN (1888 -- 'had not seen since marriage')
       STOC (1888 -- 'the June after my marriage')
       NAVA (1888 -- 'July after my marriage')
       CROO (1888 -- 'the summer of my marriage' (August))
       IDEN (1888 -- 'had not seen for some weeks')
       BOSC, TWIS (1889 -- June)
       ENGR (1889 -- summer)
       FIVE (1889 -- September)
       BLUE (1889 -- December)
       COPP (1890 -- spring?)
       REDH (1890 -- October)
       FINA (1891 -- May)

Note added July, 2004: One cannot look at the dates above without remarking on the gap between SPEC and REIG, which is something like four years (to the month!). It may (or may not) be significant that REIG begins with an exhausted Holmes, quite in need of recuperation. I have decided to mark this period (in the table which acomanies this essay) as the "lesser hiatus" (which is Watson's).

*    *    *
The A.R. cases of the Canon present an altogether different problem, for they did not begin to be published until Holmes was near retirement. This lapse was due to his own stricture (mentioned in EMPT and SIXN); we know from SIGN that he was not altogether happy about Watson's treatment of the events of STUD, and he also seems to have quashed any further publications after SIGN appeared, which prohibition endured until, of course, he himself left the scene. He would seem to have re-instituted it upon his return. As a consequence, when he began publishing again Watson had virtually the whole corpus of Holmes' work to draw upon.
    Although all appeared under Watson's name (or rather, under the nom-de-plume of his literary agent), it seems possible that he did not write all: BLAN and LION appear to have been written by Holmes himself, and MAZA and LAST would seem to have been penned by some unknown third person. They do appear, however, to be actual cases and Watson has, after all, given them his imprimatur.
    Two -- LAST and the aforementioned LION are clearly dated as being after Holmes' retirement, and may be unhesitatingly placed at the end of our chronology, in the reverse of the above order; that is to say, LAST comes last.
    HOUN is clearly a B.R. case - indeed it seems to be dated late in 1889.

Note added July, 2002: A somewhat technical study of the appearances of the moon in HOUN has led to the possibility that the events may have taken place in 1887, rather than 1889. If this is so, then the objection dealt with below does not arise, as Watson would not have been married (although he would have been engaged). Although not in the original spirit of this undertaking (indeed, the investigation seems to belong to the sink-rate of parsley school), I have elected to accept it as it eliminates one of the "where was Mary?" questions which come up from time to time.
    Note further added March 2005: The case of "The Noble Bachelor," which occurred "shortly before [Watson's] marriage" would seem to have just preceeded the excursion to the Moor, as the date on Frank Moulton's hotel bill was Oct. 4.

Although Watson is married at this time, there is no mention of his wife in the story. Since Watson had been widowed by the time it was written, he may well have recast the story to avoid mention (and memories) of his wife. It is clearly quite well written, and may represent an effort by Watson to demonstrate to Holmes that he could be trusted to do justice to Holmes' cases. It seems to have succeeded in this respect, although what Holmes thought when the new publications came out sounding quite like the ones that had preceded them is not recorded.
    VALL is also pretty clearly stated as B.R., and it seems it must be since Professor Moriarty is represented as alive -- if rather behind the scenes -- in it. This is often cited as a problem since Watson was not supposed to have heard of the Professor until FINA. The standard way 'round this problem seems to this writer to be backwards: instead of Watson concealing knowledge in FINA, it seems a good deal more probable that Watson inserted Moriarty into VALL because Holmes had later told him he had been involved. Also, by giving himself the speech about his (Moriarty's) fame among the criminal classes, he could score off a bit on Holmes. There are reasons to suspect that not all elements of the story share its 'late eighties' date, but the case is clearly B.R.
    SECO is B.R. as well, being added to the dozen tales that make up the rest of the Return to fulfill a promise made in YELL, where it is dated to "the July after my marriage."
    The majority of the remainder of the cases are given dates which place them in either the '90's or the early 1900's; the others have no explicit dates. The undated cases are:

       CHAS, SIXN, REDC, LADY, THOR, SUSS, 3GAB, SHOS

Those with dates (or inferable dates) are:

       EMPT, NORW, SOLI, DANC, PRIO, BLAC, 3STU, GOLD, MISS, ABBE,
       WIST, BRUC, DEVI, DYIN, MAZA, CREE, 3GAR, ILLU, BLAN, RETI, VEIL

(DANC comes the year after 'the' Jubilee -- only the (Diamond) one of 1897 seems reasonable; DYIN is dated to the 'second year' of Watson's marriage (the first and only, which places it in 1889); MAZA has no date, but the use of a phonograph places it late A.R., and the fact that Watson is not living at 221B places it after the time he moved out.
    The Duke of Holderness has been Lord-Lieutenant of Hallamshire since 1900, so PRIO must postdate that. Thorneycroft Huxtable collapsed onto Holmes' bearskin rug on the Thursday immediately following Monday May 13, and a check with a perpetual calendar gives 1901 as the only plausible year. [Revised July 2003]

In view of the fuss that has been made over the issue of Watson's wives, it may come as some shock to discover that only two of the above-mentioned cases contain any reference to a wife for Watson. As just noted, a reference to the 'second year of my marriage' is found in DYIN, and another occurs in BLAN, where Holmes makes his famous complaint that Watson had deserted him for a wife ("the only selfish act ..."). Now Mary Morstan seems to have died sometime between FINA and EMPT, and apart from this, no other woman has been acknowledged as taking her place. We may furthermore note that Holmes was not at all happy with Watson's announcement of his marriage in SIGN, refusing to congratulate him, and saying somewhat mysteriously "[she] might have been useful in such work as we have been doing."
    I have suggested above that Watson used the (unknown to him then) behind-the-scenes presence of Moriarty in VALL to 'score off' Holmes with a bit of 'pawky' humor, and it does not seem beyond all reason to suppose that Holmes might have taken this opportunity (for he is undoubtedly the author of BLAN) to insert a not-too-gentle reminder of his reaction to that long ago episode in their past by way of pay-back. So far as I am concerned, this disposes of the issue of multiple wives for Watson. He was married once, widowed once, and that's that.
    A more serious degree of confusion in these later cases arises over the somewhat conjoined issues of where Watson was living, and whether or not he was in practice. In NORW (which clearly takes place after EMPT -- but see below) we learn that he has sold his practice -- to a young Doctor Verner who turned out to be a relative of Holmes -- and moved back to Baker Street. However in some of the later tales he seems to be living apart and in some of these he seems to have a medical practice.
    This is perhaps not as surprising as it might at first seem, for Holmes apparently had some sense of 'mission' in his career as the world's first consulting detective, and to have been quite satisfied with the way it went, so that he, as we would now say, took early retirement. As he did so when he was fairly young (before he was 50, it would seem) it is not altogether surprising that Watson might have discovered a desire to get back into harness, as it were, before time erased his own professional skills. In addition (if addition be needed) we might remark that Holmes cannot have been the most congenial of room-mates, and his refusal (it almost has to be that) to move out of the cramped Baker Street quarters when his success must have brought more than adequate money to do so must have made things increasingly uncomfortable for Watson, who was, after all confined to a bedroom and a shared sitting-room (after having had a whole house -- and a wife -- to himself).
    As I have indicated above, the majority of the A.R. cases 'come with' dates, or indications of dates. It seems to be accepted Sherlockian tradition to automatically question any date that Watson writes down, mostly because it is occasionally found that the day of the week he gives does not correspond to his accompanying month, day and year. I submit that this is carping, and once we allow it to get a foothold, we will be reduced to questioning almost anything Watson says, turning the entirety of the Canon into nothing more than a work of fiction.

What are we to do? Well, I propose something novel (at least so it seems to me) ... as I said my goal is a simple ordering of the cases, and not an absolute fixing of their times, so I have decided to simply accept Watson's references to year and general time of year (leaving to one side anything more specific) and simply see what results. If what results is a total mishmash, devoid of all order, we shall declare this exercise a failure and go home (since you are reading this, it should be fairly clear that I did not 'go home').
    Some ambiguity will no doubt remain ... I propose to simply allow it to remain, so that what this chronology becomes is nothing more (or less) than the inherent, in-dwelling, chronology of Watson's reports. Those who wish more are free to pursue their goal(s). [And, as this is a preliminary study, it remains to examine each of these cases individually to see if there is some objection to Watson's date.]

Having said all that, what is the order of the dated A.R. cases? From the internally-supplied chronology, it is:

       EMPT (1894 -- April)
       GOLD (1894 -- November)
       WIST (1895 -- end of March)**
       SOLI (1895 -- end of April)
       3STU (1895 -- May)
       BLAC (1895 -- July)
       NORW (1895 -- August)
       BRUC (1895 -- early November)
       MISS, VEIL* (1896)
       DEVI (1897 -- March)
       ABBE (1897)
       DANC (1898 (inferred) -- July)
       RETI (1899 (inferred))
       PRIO (1901 (inferred) -- May)
       3GAR (1902 -- July)
       ILLU* (1902 -- September)
       BLAN* (1903 -- January)
       CREE* (1903 -- September)
       LION* (1907 -- July)
       LAST* (1914 -- August)

 **  Careful readers of the Canon will of course object that WIST is given as occurring in 1892; this was so up until the publication of the Oxford Sherlock Holmes, which corrected this obvious error. The matter of ex-President Murillo is mentioned in NORW, and it is suggested that Watson inadvertently dated the case to about a year after Holmes disappearance, instead of a year after his return. This seems reasonable to us, and so we have used the Oxford date.
    This creates a considerable problem for NORW, which occurs in August, and must now be moved to 1895 to "clear" WIST, which makes it occur after several other cases in which Watson seems to have already moved back in with Holmes. We can only assume that since Watson published NORW immediatly after EMPT, he of course mentioned that he had returned to 221B, but neglected to mention that more cases than WIST came before NORW.

MISS is a special case: it is stated as occurring in February "some seven or eight years ago" As it was published in 1904, this means that the earlier of its two possible dates is February 1896, before VEIL (which is "late" in 1896), and the other possibility would be February 1897, shortly after VEIL. The listing on a single line is intended to reflect this ambiguity by implying that the cases may have occurred in either order.
    I have added asterisks to those cases in which it seems that Watson may have been living apart from Holmes, after having moved back after EMPT (we know that he moved out at some time, since he says so in ILLU). In glancing over this list (having just typed it), I have to say that it seems to me a pretty reasonable sort of thing. 1895 does show up as a sort of 'banner year', and no year from then to the turn of the century is unrepresented. The 'living apart' cases are all, with the exception of VEIL, grouped in the later years, as one would expect. In fine, this does not look like the work of a man who is picking dates by throwing darts.
    There are two possibilities for the anomaly of VEIL: Watson has either gotten the date wrong, or has misremembered the circumstances. As the case was published rather late (in 1926, in fact) and evidently after some attempt to 'get at and destroy' Watson's papers, it seems likely that he had the circumstances surrounding 'the politician, the lighthouse, and the trained cormorant' in mind when he recast the events of VEIL from his notes, forgetting that in 1896 he was still residing at 221B.
    As is clear from VEIL, Watson does not pick his cases haphazardly, for many have introductory paragraphs remarking on how they came to be chosen. I do not think, however, that this means that he reviewed all of what must have been hundreds of cases each time he sat down to write a story. What I think a much more likely procedure is that he would most often have begun with the last one he wrote up, paging forward to find another one, the reviewing of the cases rejected leading to the introductory paragraphs. Equally obviously, he did not go through his notes only once, but will have occasionally gone back for a fresh start. What this means for this chronology is that an undated case might well have come in actual time between the cases it finds itself between in the order of publication. (This will merely serve to give us some place to put them, the necessary tentativeness of this 'dating' being fully appreciated.)

We are left with the internally 'undated' cases, 8 out of the 34 A.R.'s:

       CHAS, SIXN, REDC, LADY, THOR, SUSS, 3GAB, SHOS

CHAS is a good example of the problems a chronologist faces. Watson makes it very clear that he is obscuring times, places, and names (and well he should, since he and Holmes are concealing a murder). The case was published in 1904, and the only direct clue as to time is Watson's opening remark that "It is years since the incidents ... took place ..." As 1904 is only ten years removed in time from 1894 -- the year Holmes resumed his practice -- I think it is not unreasonable to assume that it occurred in Holmes' earlier career (B.R.) and to look to the only other clues the story provides, namely that Watson seems to be residing with Holmes at the time, and Holmes seems to be young enough to pass himself off as a 'rising plumber' -- something a fortyish man could hardly expect to be successful at. Thus this case really does have to be moved to B.R. status, and ante-marriage status to boot. This gives us the block of time to which it belongs, but no way to fix it within that span (seeing its precedence over SCAN, however, does give us some insight into Watson's remark in that story that when it comes to breaking the law, "I am your man"). Watson's ability to scramble up a wall and run across Hampstead Heath suggests that the events took place long after his Afghan mishap, perhaps around the time of HOUN in which he also seems to have recovered his 'fleetness of foot'.
    The above gives us, I would say, as good a sense of when CHAS occurred as we can hope to have, and if we are unable even to assign it to a year, well, so be it: I quite expect that many of us who are approaching our Reichenbach have many memories that suffer the same defect.
    SIXN also has no internal date, but as it follows fairly closely upon the first few cases of the Return, and as Lestrade delivers himself of a veritable encomium of praise for Holmes, one might suppose that the case might have occurred not long after Holmes' return from the dead. (Nothing like a resurrection to enhance your image.) As it was published amongst many other 1895 cases, I was initially inclined to place it shortly after 3STU and before BLAC, where there seemed to be an inviting gap in the 1895 casework. But this is at odds with the impression created by Lestrade's visit which opens the case (Watson says "it was no very unusual thing for [him] to look in ... of an evening". This seems to suggest a practice of some long standing, and not one recently initiated. These notions are in conflict, and absent more specific indications, we shall have to leave the year of this case unassigned.
    REDC has only a few internal clues as to its time, and these are highly inferential. First, Watson is clearly living with Holmes, and Holmes makes a remark about "when you were doctoring". This rather fixes the time to between when Watson's practice was bought out (before NORW) to when he clearly had resumed it. Also, this was a journey that ended in lovers' meetings -- or so Holmes said when he and Gregson collided on Howe Street. As we first met Gregson in STUD, and he still seems to be a detective (despite being the smartest of them), we may fairly, I think, leave this case about where Watson placed it, which is to say after DEVI.
    LADY is also a case with few self-contained clues, beyond the inference that Watson is living with Holmes and has no difficulties with setting of for the continent on short notice. We shall also simply leave it where Watson published it.
    THOR is even less forthcoming with hints as to its date (although we find it is in October). Although we do know Watson is living with Holmes, it, too, remains near where Watson put it, but moved from before CREE to shortly after.
    In SUSS we have only Watson's remarks of the effects of the passage of time on 'Bob' Fergusson (and himself) to guide us, but these do suggest that this is truly an A.R. case.
    3GAB was published immediately after ILLU (but with some lapse of time), and it does seem that Watson has moved out, although his language is not explicit. He speaks of not seeing Holmes 'for some days', and of being 'settled ... into the well-worn low arm-chair', from which we may fairly infer that he is now a visitor to the rooms he once called home. However, there seem to be no clues as to the time of year, and so we do not know if it took place in 1902 (the year of ILLU) or later. It thus can be placed only as an A.R. case when Watson was living apart from Holmes.
    SHOS, the last story which Watson published, is another problem for chronologists, for it has no indications of time but requires subtle interpretations of subtle clues. It may well be another B.R. case, for Watson speaks of his 'wound pension' and spending half of it on the ponies. 'Pension' has connotations of a lasting benefit, but wound pensions were evidently reviewed periodically and discontinued if the disability incurred had ceased. Watson's 'wound pension' is likely not to have been due to his (wandering) wound, but to the lasting aftereffects of the enteric fever which almost took his life. It was this which (so he thought at the time) "irretrievably ruined" his health and gravitated him to London where he took up residence with Holmes. A casual remark in the story suggests that the case occurred early in their acquaintance. As Holmes is examining some evidence under the microscope, Watson asks "Is this one of your cases then?" This phrasing suggests that Watson is still in the finding-out-about-his-roomate phase of their acquaintance, and Holmes' reply that he is "looking into it for Merivale of the Yard" suggests that Holmes is still in the consulting-on-minor-matters phase of his career. When we consider that Watson seems to have lost his knowledge of racing by the time of SILV, we can, I think, with some confidence, move this case to B.R. status, and early in that time frame to boot.

*    *    *
In seeking a way to display all of the cases, it occurred to me that it would be fruitful to look back to the topological approach with which I began. Holmes' cases fall readily into 3 categories: early (pre-Watson), middle, and late (post-retirement). There are two early (GLOR and MUSG) and two late (LION and LAST). The middle cases are divided by two cases which cover the episode of the Reichenbach (FINA and EMPT). Each of the resulting groups is also subdivided: the earlier ones into pre- and post-marriage, and later ones into pre- and post-moving-out/resuming practice. If we use these categories to give an over-all structure, we can construct the following table which summarizes these findings. Note that the (published) A.R. cases determined to have actually taken place B.R. (e.g. HOUN) have been inserted into the previous B.R. listings.

Canonical Chronology

Time division	Ordered cases	Grouped cases
Pre-Watson	GLOR (1874 -- summer) MUSG (1878 -- July)	.
Before marriage	STUD (1881 -- March) RESI (1881 -- October) CARD (1882 -- August) SPEC (1883 -- April) REIG (1887 -- April) GREE (1887 -- summer) SIGN (1887 -- September) NOBL (few weeks before m. -- early Oct.) HOUN (1887 -- mid October)	BERY SILV YELL SHOS CHAS
After marriage	SCAN (1888 -- March) STOC (1888 -- June) NAVA, SECO (1888 -- July) CROO (1888 -- August) IDEN (1888 -- some weeks after CROO) BOSC, TWIS (1889 -- June) ENGR (1889 -- summer) FIVE (1889 -- late September) DYIN (1889 -- November) BLUE (1889 -- December) COPP (1890 -- early spring) REDH (1890 -- October)	VALL
Great Hiatus	FINA (1891 -- May) EMPT (1894 -- April)	.
Back at Baker Street	GOLD (1894 -- November) WIST (1895 -- March) SOLI (1895 -- April) 3STU (1895 -- May) BLAC (1895 -- July) NORW (1895 -- August) BRUC (1895 -- November) VEIL (1896 -- late in year) MISS (1896/7 -- February) DEVI (1897 -- March) ABBE (1897) DANC (1898 (inferred) -- July) RETI (1899 (inferred)) PRIO (1901 (inferred) -- May) 3GAR (1902 -- July)	SIXN REDC LADY THOR SUSS
Independent living	ILLU (1902 -- September) BLAN (1903 -- January) CREE (1903 -- September)	MAZA 3GAB
Retirement	LION (1907 -- July) LAST (1914 -- August)	.

UNDERLINING INDICATES B.R. CASES PUBLISHED A.R.
Italic indicates date changed from Canon

To print just the table, use this page.

This table gives us, I would submit, the basic chronological information we can wrest from the Canon. Some 14 cases -- less than a quarter of the Canon -- have no specific year, but can be placed with respect to the major divisions of Holmes and Watson's joint career. It should be noted that the groups of the undated cases in the table are not ordered; they must, of course be written down in some order, but nothing is intended by that. Dated cases listed on the same line are similarly unordered.
    An alternative way of presenting this data has occured to me, and I append a new page to show it. It lists the cases, book by Oxford Sherlock Holmes book against the passing years. You will find under each book those stories which appear in it, but arranged in the way they actually occured (at least according to this chronology).     A curious thing about this chronology is its oddly symmetric structure -- hardly intended and indeed not noticed until it was finished. First, there is the 'blocking off' of the two main phases of Holmes' career (before and after the Moriarty incident) by pairs of stories: the two pre-Watson stories, the two post-retirement stories, and the pair which deal with the Moriarty episode. Second, the remaining 50 short accounts are distributed equally between the two time periods, 25 B.R. and 25 A.R. Lastly, while the longer accounts are unsymmetrically all found in the B.R. time frame, they do divide two by two before and after Watson's marriage (but not if we move HOUN to 1887). No significance would seem to attach to these features, but it is amusing to note them.
    Obviously efforts to further refine this table are quite possible, and indeed the author himself might even undertake some (as he has -- cf. the change to the date of HOUN). It should further be noted that this somewhat straight-forward examination has -- in this author's view -- effectively quashed the question of multiple wives for Watson. (It is, of course, quite possible to consider that "independent living" means "took a second wife", one who set down as a condition of matrimony that she be neither named nor discussed in any additional cases Watson wrote up. It is possible, but this does not constitute a reason to suppose it, and, following the injunction of Ockham's Razor not to multiply entities (in this case, wives) beyond necessity, I choose not to.)

What we have here -- the end result -- is not quite what was envisioned when this exercise was begun. Rather than a single, ordered list we have a dual categorization in which some cases are ordered (with some ambiguities of order remaining), and some are located only in terms of the most general periods of time describing Holmes' career. As I accepted uncritically almost all of the dates which Watson provided (only two were changed) -- and was rather pleased to see that no chronos horribilis resulted from this -- what we would seem to have is what might be termed the base chronology of the Canon: its prima facie chronology, if you will. I use the term 'base' with some hesitation, not because of its more unfortunate connotations, but because it seems to bestow a position of some eminence upon what is, after all, a somewhat casual effort. No such claim in intended.
    Although this chronology did not reach its stated goal of placing all the cases in their temporal order, it has come close enough to raise the specter of a possible publication in its own order, after the fashion of the famous Baring-Gould annotated edition of the Canon. It should be said at the outset that in the discussion that follows, what is being envisioned is not another annotated edition. That need is being amply and masterfully met by the currently appearing Sherlock Holmes Reference Library editions of Leslie Klinger. What we imagine is more on the order of a reading edition ... read more

Some Observations
The slight use made of the dates of publication of the cases in this chronology has brought these dates into focus with regard to the main events of the Holmes-Watson collaboration, and as a result a most interesting pattern has emerged that, I believe says a good deal about the relation between Holmes and his "Boswell". Let us mention the main elements of this pattern:

Upon the publication of A Study in Scarlet, the first case which Watson chronicled, Watson married and moved out.

The Sign of the Four was published in England in early 1890, when Holmes had begun his famous drawing of toils about Professor Moriiarty, and was, or so Watson thought, killed a year or so later. A Scandal in Bohemia was published two months after Holmes' demise, and all of the "B.R." short cases were published in the next two years.

Holmes returned from the "Great Hiatus" in early 1894, and Watson's publications immediately ceased until 1901 when the serial publication of The Hound of the Baskervilles began. Shortly thereafter (in 1902) Watson moved out again (Holmes hints that he may have taken a wife, but Watson does not confirm this).

In 1903, Holmes retired, to keep bees upon the Sussex Downs, and Watson published another spate of short cases, covering primarily ones which took place during Watson's second tenure at 221B..

The remaining elements of the Canon were published at irregular intervals from 1908 to 1927.

The singular feature disclosed in this summary is that virtually the whole of the Canon seems to have been written (and certainly published) at times when Watson was not at Baker Street, and not in regular contact with Holmes. We know from Watson's remark in The Empty House that "[o]nly now, at the end of nearly ten years," was he "allowed to supply the missing links" to that remarkable case. It would seem that stricture applied to any publication of Holmes cases, whether past or current, as Holmes says, in The Six Napoleons, that he is sure Watson will "enliven [his] pages with an account [of that adventure] "[i]f ever I permit you to chronicle any more of my little problems ..." We may fairly ask why such a prohibition might have been issued.
    A hint is contained in the decidedly frosty reception which the publication of A Study in Scarlet seems to have received (as recorded by Watson in The Sign of the Four), and there are the occasional remarks inserted passim about Watson's method of telling stories ("backward"). Why Holmes should have felt this way is not at all clear, but the publication of The Hound of the Baskervilles seems to have been something of a breach of this agreement, which might explain Watson's move to other quarters.
    However, upon his retirement, Holmes does seem to have released Watson fully from this stricture, for we have another period of rapid publication following upon it. Holmes seems to have either mellowed or resigned himself to Watson's penchant for publication, for not too many years later we find him going so far as to suggest a case for Watson to publsh ("Why not tell them of the Cornish Horror ...").
    When we look at the actual dates when the stories were published, it becomes obvious that we must take with a grain of salt (perhaps even a quintal?) remarks inserted into these tales which imply a near-contemporaneous publication -- the most famous being Mycroft's remark (in The Greek Interpreter) that "We hear of Sherlock everywhere now that you have become his chronicler.", for at that time Watson had published only one case. This could, of course, have been said tongue-in-cheek; a man who could conceive of the Diogenes Club has to have had some sense of humor. Indeed the remark often most taken to imply that Watson was publicising Holmes -- "I am lost without my Boswell" (uttered in Scandal) -- is probably misconstrued if so taken, for Boswell wrote his Life of Johnson after Johnson has passed beyond his own Reichenbach, just as Watson thought Holmes had when he began writing up the short cases. And it is clear from many stories that Holmes did not lack for coverage in the contemporary press. (The most obvious example would be in The Reigate Squire, where Watson found him in "the blackest depression" despite the fact that at that "time Europe was ringing with his name" and "his room was literally ankle-deep in congratulatory telegrams.") It is to the other aspect of the Boswellian connection that we must turn to find its meaning: the recording by Boswell of a great friendship. This may be the closest the reticent Holmes could come to saying how much he valued Watson's companionship.

Copyright 2000-2004 David Richardson