In 1994 the University of Toronto Press granted me permission to put on-line its out-of-print undergraduate poetry textbook, Representative Poetry. First edited in 1912 by Professor W.J. Alexander, whose name honours an annual lecture series at University College, Toronto, the collection then sold for one dollar. When the book re-appeared in 1916, W.H. Clawson joined Alexander as co-editor. Their 422 pages included works by 69 poets from Sir Thomas Wyatt to William Morris. This collection proved useful and became a 50-year tradition in the Toronto English department. The next important revision, published in 1935 and substantially re-issued with only minor corrections until 1960, was edited by seven members of the department, led by Clawson and R.S. Knox. Two of these seven were still with the department when I arrived at Toronto as an M.A. student from Manitoba in 1964: N.J. Endicott and J.R. MacGillivray. The two-volume 1935 edition included works by 152 poets from John Gower to Algernon Charles Swinburne, over twice the number found in the 1916 edition, and had 1454 pages. By 1960, the combined departments of English, led by editors F.E.L. Priestley from University College and F.D. Hoeniger from Victoria College, undertook the fourth major revision of the collection. They were assisted by 17 colleagues including Endicott and MacGillivray. This stellar field encompassed, for example, Kathleen Coburn (for S.T. Coleridge), Northrop Frye (for Christopher Smart and William Blake), Milton Wilson (for lord Byron and Percy Bysshe Shelley), A.S.P. Woodhouse and Hugh MacCallum (for John Milton), and Marshall McLuhan (for Alfred lord Tennyson). The three-volume text had only 23 fewer pages than in 1935 but covered barely a third of the poets, just 56. I used this edition, the last printed one (termed "the third"), to teach a course from Chaucer to Milton when I joined the department in 1968. A few years later the combined departments of English at Toronto decided against continuing their work on Representative Poetry, partly because a new required curriculum had done away with the three survey courses that employed the three volumes, partly because faculty wanted to teach from textbooks of their own choosing. An account of the series -- the first book to come out from the University of Toronto Press -- appears in Robin Harris' history of the Toronto English department (Harris 1988).

The first World Wide Web edition or version went on-line December 15, 1994. It was based on the last printed edition (1962-63), with new poems by Elizabeth I, Aemilia Lanyer, Anne Killigrew, and Thomas Randolph. Sharine Leung at the Toronto Centre for Computing in the Humanities produced rough-scanned copy of the three volumes, Sian Meikle of the university library and I co-designed the Web page, and Meikle also enabled a simple keyword-search of the collection. I did the proofreading, editing, encoding, indexing, and file management. Rebuilding Representative Poetry for the Web meant splitting the collection into as many files as there were poems. Each poem had to be keyed to a table of contents file for the poet concerned, and to an on-line source bibliography in which entries, where possible, were identified by local library name and shelfmark. I used a version of emacs as text-editor, perl to write line-numbering and filtering programs, sed to transform my COCOA- and SGML-encoded files into HTML documents, and grep to extract information for indexes by author, title, first line, and date. Notes to the poems (except for the bibliographical references) were not included because they were still copyrighted to the editors and will remain so for another 20 years.

By creating Representative Poetry On-line we were able to provide instructors and students of English poetry with good, free reading editions on a very new medium, the World Wide Web. These could also be searched, concorded, and processed variously by text-analysis tools, once available electronically. Believing poetry to be a civilizing presence in human life, and seeing the Internet awash with much less worthwhile reading, I thought that English faculty, especially ones like myself with a history in computing in the humanities, had a role to play in putting great literature on-line.

I placed my e-mail address in the Web edition so as to find out what readers thought of the collection. They confirmed the views of the General Editors in 1962-63 that the anthology serves "students, particularly in the pass (now the general) course", as well as "the ordinary reader". Students undertaking assignments, poetry lovers trying to identify half-remembered lines, and readers looking for spiritual comfort or for poems to fit occasions have written me from many places, including Argentina, Bangladore, Spain, Sweden, and Germany. Some catch proofreading slips, ask about variant readings, and complain about omissions. A Brown University student angrily denounced the collection to me as non-representative because it lacked black American poets. Colleagues at home politely wondered where the native poets, the women poets, the Catholic poets, and the Canadians were. However, people were generally grateful and complimentary because, apart from the Chadwick-Healey poetry database, this is the largest collection of its kind; and it is free. Other sites now link to Representative Poetry On-line.

In 1995 I began making the collection more representative of early and late 20th-century critical opinion about fine English poetry. First I added nine more late 19th- and early 20th-century poets (including two Catholics, four poets of World War I, and one Canadian): the new ones were Rupert Brooke, Lewis Carroll, Ernest Dowson, Gerard Manley Hopkins, A.E. Housman, John McCrae, Wilfrid Owen, Isaac Rosenberg, and Edward Thomas. Resorting only to facsimiles of manuscripts and early editions for copytext, and rereading the poets' complete works to form my own judgments, the editing took more time than I had estimated. Legal considerations, which prevented the original editors from including their works, also constrained me from going further into our century. (British copyright law then protected an author's works for fifty years, plus war years, from her or his death date. For example, W.B. Yeats died in 1939, T.S. Eliot in 1965, and W.H. Auden in 1973, so that their works are all protected until 1996, 2022, and 2030, respectively.) I had also partly edited works by many women poets, as well as hundreds of poems from the 1935 edition that were left out of the 1960s edition. By early August, although only the first eight poets were ready, there was enough new to make a new version: revised indexes and bibliography, a "What's New" file, corrections in formatting, and, for the first time, notes to Arnold and Meredith contributed by Harvey Kerpneck, one of the original editors of the press edition. Shannon Turlington's Walking the World Wide Web, 2nd ed., now lists this second version as a "Premiere Site" (http://www.vmedia.com/books/). It is also released in the Webdoc series by the Research Libraries Group.

The present, or third version of Representative Poetry On-line (1996) adds works from the 1935 edition by over 100 poets not retained in the 1960s edition. Many are minor but well-known authors in their periods. The 1960s edition failed to carry forward some fine poems in the preceding edition: John Dryden's "Alexander's Feast", Shakespeare's Sonnet 97, "How like a Winter hath my Absence been", Shelley's "Ozimandias", Tennyson's "Break, Break, Break", "Crossing the Bar", and poem 54 from In Memoriam, "Oh, yet we Trust that somehow Good". One reader asked for "Ozimandias". Four other types of omissions point to selection criteria by the original editors that would be questionable today: the absence of women poets and of scandalous, popular (especially children's) and pre-Renaissance writers. For that reason I collected and edited 86 more poems by some 27 poets, in addition to providing the Hengwrt manuscript of Geoffrey Chaucer's "Canterbury Tales", Edmund Spenser's "Faerie Queene" (Book I), and the complete "Lyrical Ballads" by William Wordsworth and S.T. Coleridge.

Twenty of the 27 new poets are women: Mary Barber, Aphra Behn, Anne, Charlotte, and Emily Brontë, Elizabeth Barrett Browning, Lady Mary Chudleigh, Mary Coleridge, Eliza Cook, the Canadian Mohawk Pauline Johnson, Alice Meynell, Mary Monck, Lady Mary Wortley Montagu, Amelia Opie, Katherine Philips, Adelaide A. Procter, Mary Robinson, Charlotte Smith, and Ann and Jane Taylor. Of these, Barrett Browning and Emily Brontë stand up well to any poet of the 19th century. Aphra Behn's poem, "The Disappointment", about a young man's inability to achieve an erection just when he and the maid he has seduced are expecting one, is one of the most striking erotic poems of any century. I added this poem but have had doubts about the brilliant, obscene erotica by John Wilmot, Earl of Rochester. These are still not included. Under what criterion can poems filled with four-letter words, and overflowing with contempt for women, be included in this collection? These poems are fine, but also plainly offensive to many good and intelligent people. The opinions of colleagues I have canvassed so far agree in recommending their exclusion. Popular and children's literature, in contrast, offered few problems. One correspondent, for example, asked about Robert Service. Joining "The Shooting of Dan McGrew" and three other poems by Service are "The Hunting of the Snark" by Lewis Carroll, the childlike and often sad lyrics of Eliza Cook ("The Old Arm-Chair"), Amelia Opie ("Yes, Mary Ann, I freely grant"), and Ann and Jane Taylor ("Twinkle, Twinkle, Little Star"), and the sterner stuff of Robert Louis Stevenson's "Child's Garden". Finally, I added later poets in copyright earlier, but now in the public domain: Austin Dobson, Ernest Henley, Alice Meynell, Francis Thompson, and Oscar Wilde (his superb "The Ballad of Reading Gaol").

In fall 1996, Representative Poetry On-line has just about 1350 poems by 207 poets. A revision of the British copyright law passed in January 1996 caused me at first to exclude A.E. Housman. This change extended the period of copyright from 50 years (plus war years) to 70 years after the author's death. The works of Housman, who died in 1936, consequently will come back into copyright once the law takes effect. Because in-progress editions are not affected by the law, however, Housman has been restored to the collection. Soon works by poets who died after 1926 will be circulated on-line only by persons who break copyright.

The World Wide Web collection supplies the core of a hypertext poetry corpus for teaching purposes. Students can search any computerized corpus quickly for keywords and references and so supplement the critical thinking that can only come from reading. By exploring electronic text collections like Representative Poetry On-line, students can also intelligently increase the number of texts they choose to read for essays and other assignments. The pedagogic advantage for teachers -- over commercial databases and even the holdings of the Internet itself -- is flexibility in adding to, deleting from, or modifying the texts of the collection as well as the power to make links between the poems and among them and any specialized indexes.

For example, the recent on-line version gives an index of prose and verse criticism on poetry that ties together twenty newly-added works of prose criticism and a large number of poems in the collection that concern poets and the writing of poetry. The prose works extend over three centuries. The Arte of English Poesie, attributed to George Puttenham (1589), surveys most of the rhetorical devices imported into English verse from classical literature. Sir Philip Sidney's Defence of Poesie (1595), John Dryden's Of Dramatic Poesie (1668), Edward Young's Conjectures on Original Composition (1759), the Preface to Lyrical Ballads by William Wordsworth and Samuel Taylor Coleridge (1798), Thomas Love Peacock's The Four Ages of Poetry and Percy Bysshe Shelley's reply to it, Defence of Poetry (1821), take a variety of perspectives on the value and importance of poetry, from the Renaissance to the Romantic period. The theory of criticism of poetry can be traced from Ben Jonson's Timber (1640), through John Dennis' The Grounds of Criticism in Poetry (1704) and the first volume of Coleridge's Biographia Literaria (1817), to Matthew Arnold's "The Function of Criticism at the Present Time" (1865). Accounts of individual writers appear in works like Thomas Spratt's biography of Abraham Cowley (1668), Pope's preface to The Iliad (1715), Samuel Johnson's preface to Shakespeare (1765), and Walter Pater's essays on Coleridge and Wordsworth. The verse criticism includes over 50 poems, classified under the names of 35 poets of whom they treat, and under topics such as "Advice to Future Poets", "Bad Poets", "Critics and Criticism", "Fancy and Imagination", and "The Poet's Life and Making".

These works illustrate how the English poetic tradition, what we used to call the canon, emerged over four centuries. Poets, later joined by poet-critics, changed the type of poetry that was being written by assembling an earlier library of great poets and poems. Later poets tended to emulate or comment on their predecessors in verse.

An on-line library such as this can also help study the criteria that modern critics use in judging poetry as typical of its period, or as especially good. Representative Poetry On-line consists of three chronological strata of texts, each reflecting critical taste at a different period, in 1935, 1962, and today, but I will here refer to the 1912 and 1916 editions as well. With respect to the 18th century, for instance, critical interest in Robert Burns, Thomas Gray, and Alexander Pope has remained steady, and the importance of John Dryden and especially William Blake has increased, but Thomas Parnell, Matthew Prior, John Wilmot, Earl of Rochester, and Edward Young disappeared from the 1960s edition -- in fact 61% of the poets in 1935-36 vanished. Anne Finch, John Gay, Samuel Johnson, and Christopher Smart appeared only after 1916, and Thomas Percy's Reliques in 1962. Between the 1930s and the 1960s the editors dropped dozens of poets only represented by one or two poems and chose to increase the average of poems per poet from 3.7 to 6. By conflating all editions electronically into one Web database, we can begin to see how what we perceive of as poetry changes from generation to generation. In the 1930s the Toronto department taught poems, the more varied the better, but by 1962 it taught authors, the more fully represented the better. The author-centred canon that the editors of Representative Poetry gradually came to be seen as disintegrating today. Out in the World Wide Web where the ordinary reader dominates, a new critical consensus seems to be emerging, one that chooses poems by their subject matter and memorable lines as much as by their authors, canonical or not.

By computerizing printed volumes of poetry on a systematic basis rather than by selecting works for inclusion, editors give all poems the same value. Nothing is representative because the rationale for inclusion is bibliographical. In contrast, Representative Poetry On-line has a unity that arises from the history of critical opinion and from the expressed needs of the on-line readership. When I select for the collection, I put my own tastes on the line, although I also pay attention to what others place on-line (for example, Thomas Dell, the organizer of the wonderful Internet Wiretap, and Jeffery Triggs, the Director of the North American Reading Program for the Oxford English Dictionary). Often I rely on my intuition.

The editor-teachers who assembled this little universe did not explain their principles of inclusion, but the thousands of regularities, clusters of repeating elements, and cross references in the texts speak for themselves. Critical intuition, working unconsciously, selects poems that are part of a network we call the literary tradition. This network has innumerable strands that its hundreds of voices follow. A concordance of almost any word will elicit these literary strands.

For this reason the texts of the latest Representative Poetry On-line will also come out in the TACT CD-ROM published by the Modern Language Association this year (Lancashire 1996). This publication gave me the opportunity to do what my predecessors, the editors of the printed Representative Poetry, would certainly have liked to do: to add full texts instead of extracts, to multiply the number of poets, and to place the collection in a larger library of prose fiction, non-fiction, and plays.

Fellow scholars agreed to donate much more material to the CD-ROM and made its library of English exemplary texts more representative. The North American Reading Program for the Oxford English Dictionary alone, directed by Jeffery Triggs, contributed 250 works. Medieval poetry, not represented in the third edition of Representative Poetry at all, was supplied by Karen Arthur, Francis McSparran, Stephen Reimer, John Leyerle, Norman Blake, David Burnley, the Berkeley SunSite Medieval and Renaissance Library series, and Gary Shawver. For the Renaissance period, Richard Bear contributed all of Spenser's works and much of Sidney's, Raymond Siemens a new edition of Tottels Miscellany, Hardy Cook Shakespeare's Venus and Adonis (and both of us the sonnets), Donald Foster the newly-attributed Funeral Elegy, and Roy Flannagan all Milton's Paradise Lost. Jack Lynch of the University of Pennsylvania, David Mikosz, and Tom Keever added to the collection's holdings of 18th-century literature. The CD-ROM will hold more than 600 works by about 330 authors in English.

It includes electronic texts of all poems in Representative Poetry On-line, entered in author- and title-subdirectories that reflect the headings and subheadings of the International MLA Bibliography, and also collected into two text files for poets to 1700 and poets 1700-1918. Each file has its TACT textual database, which with the Usebase program becomes an interactive concordance.

Take, for example, a concordance of the word "knot.*" (see Figure 1). This obtains about 30 occurrences by 22 poets from Wyatt to Morris. Why "knot"? I simply went to K in TACT's complete word-list and then took a random walk through its keywords. Intuition led me to choose the image of the knot. It seems to encapulate the paradoxes in our lives. Wyatt uses the word for the lover's paralysis of "hope or dread". Spenser pictures spiritual death in the knots of Error's tail, Milton contrasts the flowers of Paradise with the "curious knots" of foolish art, and Burns echoes both in the "mystic knots" of his "Address to the Devil". Donne's famous phrase, "That subtle knot which makes us man", centres the image in the Christian's dilemma of body and soul, a knot only to be broken by God, in a way Spenser, Milton, and Burns would have approved. By the 19th century, the image remained, but now expressing the unifying power of nature. Consider the primrose within its knot of leaves or the knots of grass by which a young man's handhold keeps him from falling (both Wordsworth), the bunches of knot-grass in a carven imagery (Keats), the knotty arms and roots of trees embracing other trees and sustaining mountains (Shelley), and "the knots / That held the pear to the gable-wall" (Tennyson). To Tennyson these are "knots of Paradise", but to Fitzgerald they represent the old horror as he speaks of unravelling "many Knots ... by the Road; / But not the Knot of Human Death and Fate". Two other 19th-century poets, however, best illustrate how new individual talents over time add the strings that link works and thus create the literary tradition. Coleridge repeats a commonplace by Campion and Fulke Greville, "true love's knot", but he links it to none other than Donne, wreathing "iron pokers into true-love knots". Coleridge is muttering against fancy here, but he also is bringing to consciousness a submerged debate within poetry. Second, Browning's Fra Lippo Lippi lets himself down "a dozen knots" from a room painting saints into Florence's night life. In doing so he summarizes the symbolic history of the image.

Today the World Wide Web frees us from the constraints of printed textbooks and presents literary tradition as many generations have seen it. Representative Poetry On-line is only one of a number of on-line collections of poetry. The ties that bind its historical critical consensus are HTML anchor tags. For the "ordinary reader", however, the "canon" of English poetry is now being recreated on-line by thousands of people. It includes little-known poets who self-publish on-line as well as the likes of Bob Dylan and Bruce Springsteen, whose lyrics find a place in well-maintained private Web sites because of the loyalty of fans. For most Web readers, Representative Poetry On-line is only part of a much larger universe of poetry.

Someone described the Toronto anthology to me as a florilegium, a collection of flowers. If by this is meant, the best and most lovely of a certain kind for the pleasure and instruction of readers, the word is well chosen. To my mind, Representative Poetry On-line is also a research tool for understanding what our literary traditions are and how we and our professional ancestors have formed them. It is a literary library of works handpicked by its own authors and their readers over many centuries.

Bibliography

• HARRIS, Robin S. (1988). English Studies at Toronto: A History, Foreword by H. Northrop Frye, University of Toronto: Governing Council.
• LANCASHIRE, Ian (1996). Using TACT with Electronic Texts: Text-Analysis Computing Tools 2.1 for MS-DOS and PC DOS, in collaboration with John Bradley, Willard McCarty, Michael Stairs & T.R. Wooldridge, New York: Modern Language Association.
• Representative Poetry (1916). Selected and Annotated for Class Use in the University of Toronto [by W.J. Alexander and W.H. Clawson.], Second and enlarged edition, University of Toronto Press, 422 pp., 69 poets (no women poets).
• Representative Poetry (1935). Selected and Annotated for use in the University of Toronto, Edited by R.S. Knox, W.H. Clawson, H.J. Davis, N.J. Endicott, J.F. Macdonald, J.D. Robins & J.R. MacGillivray, 2 vols., New edition, revised and enlarged, Toronto: University of Toronto Press, 1454 pp., 152 poets (6 women poets); repr. 1941, 1946.
• Representative Poetry (1962-3). Prepared by members of the Department of English at the University of Toronto, General editor: F.E.L. Priestley (and Dryden and Browning), Associate General Editor: F.D. Hoeniger (and 16th-century poets and Jonson), Contributing editors: Kathleen Coburn (Coleridge), N.J. Endicott (Donne and 17th-century poets), G.G. Falle (minor Restoration and 18th-century poets), Sister St. Francis (FitzGerald, Clough, the Rossettis), Northrop Frye (Smart, Blake), Douglas Grant (Dryden), H. Kerpneck (Arnold, Meredith), Hugh MacCallum (Milton), J.R. MacGillivray (Wordsworth, Keats), Millar MacLure (Spenser), H.M. McLuhan (Tennyson), P.F. Morgan (Scott, Landor, Morris, Swinburne), O.H.T. Rudzik (Clare), D.F. Theall (Pope), M.T. Wilson (Byron, Shelley), A.S.P. Woodhouse (Milton), and R.S. Woof (Coleridge), 3 vols., 3rd edition, Toronto: University of Toronto Press, 1431 pp., 56 poets (2 women poets); repr. with corrections 1967.
• Representative Poetry On-line (1994-). Edited by Ian Lancashire, Toronto: Department of English and the University of Toronto Press; 3rd version, 1996, 207 poets (25 women poets).