Reprinting of first-wave Gothic novels in the 19th century: method and dataset

For the upcoming collection Penny Dreadfuls and the Gothic which will be published at the University of Wales Press (UWP), I wanted to explore the Gothic reprints in my database of penny bloods, Price One Penny. I teamed up with Hannah-Freya Blake and ended up looking for Gothic reprints not just in penny bloods, but also in other 19^th-century collections and publishers’ series.

We found that the earlier reprints followed the current Gothic canon in their selections, though with some notable exceptions. The later ones saw the popularity of Ann Radcliffe diminish against a rising set of so-called “circulating-library bestsellers”. These three now sit half-way in and half-way out of the Gothic canon: Regina Maria Roche’s The Children of the Abbey and Elizabeth Helme’s The Farmer of Inglewood and St. Clair of the Isles.

This data paper presents the method and data supporting these findings to ensure that they can be properly reviewed, reproduced, and also reused in other projects. It contains three datasets.

Dataset A

The final table of 13 collections, series, or publishers from the 19^th century reprinting 29 first-wave Gothic novels grouped into categories.

Dataset B

A preliminary version of the previous table containing 48 publishers, collections, or series from the 19^th century that have reprinted 35 novels that could have been construed as part of first-wave Gothic.

Dataset C

A different table of 12 “libraries” present in two catalogues for Milner & Co (four in the first, eight in the second) and reprinting 15 Gothic novels, forming a subset of the 35 novels in the previous dataset.

They are available for download on GitHub, Zenodo, and Figshare.

Reuse potential

The chapter Hannah-Freya Blake and I have written covers much of the literary history that can be extracted from dataset A, but its data visualization potential certainly has not been fully realized.

Dataset B opens up many new avenues for research into publishers and series that have reprinted less first-wave Gothic novels and therefore were excluded from dataset A. Reprinting and series publishing are under-researched areas, especially in the first half of the 19^th century. Though the series present in dataset B but not in dataset A were less pertinent for Gothic studies, they might be important in other fields of book history and literature, with contents that might lean more heavily into other genres or sources, such as American novels.

Dataset C is the first bibliographic inquiry into Milner’s series. The Halifax publishing firm is best known for its “Cottage Library”, but it published many more series in varying cheap formats with the same novels appearing in multiple series. By the 1890s, it was even publishing two “Cottage Library” series, a second one being titled “Cottage Library (Pocket Edition)”. Dataset C has started the work of transcribing the contents of two digitized catalogues that could form the basis of a larger inquiry into the cheap publishing industry of the North of England in the second half of the century.

Method: Collecting the data

The data was collected in two stages: first all the reprint series in datasets A and B, then the data available on Milner’s series, presented in dataset C.

Reprint series

Dataset A is the final result of a seven-step process. Dataset B is the preliminary result after Step 4.

Step 1. Identifying Gothic reprints in penny weekly numbers

In Price One Penny, source texts for penny bloods are stored in a table entitled hypotexts, following terminology proposed by Gérard Genette in Palimpsestes (1982). It combines source texts for translations, adaptations, and reprints alike. I thus queried the database to find all source texts published in the United Kingdom before the beginning of the Victorian period in 1837.

SELECT * FROM hypotexts

WHERE

(place = "London" OR place = "Dublin" OR place = "Edinburgh" OR place = "Salisbury")

AND

(year < 1837 OR LEFT (start_date, 4) < 1837)

ORDER BY year, start_date;

It returned 107 results that I transferred into a spreadsheet in OpenOffice in the OpenDocument Spreadsheet (ODS) format. Hannah-Freya sifted through them manually to identify the Gothic novels among them.

She drew on numerous Gothic bibliographic sources to cross-reference with this Price One Penny data. These references include Garside et al’s comprehensive The English Novel 1770–1829: A Bibliographical Survey of Prose Fiction Published in the British Isles, Gary Kelly’s Varieties of Female Gothic, and the appendix of Gothic novels featured in Franz Potter’s The History of Gothic Publishing, 1800-1835.

Step 2. Examining the penny weekly publishers reprinting Gothic novels

Looking up in which penny bloods these novels were reprinted revealed the importance of the Novel Newspaper and, more surprisingly, of “T. Paine”. The Novel Newspaper is a serial anthology well known for reprinting out-of-copyright works, therefore from the United States, or issued in the United Kingdom in the 18^th century or in the 1800s, provided that the author was dead by the time of reprinting (to comply with the Copyright Act 1814). The presence of Gothic novels in its pages was nothing surprising.

Closer inspection of the publications under T. Paine’s imprint revealed that it contained at least one Gothic novel, Friar Hildargo, which had been retitled Love and Crime, or The Mystery of the Convent. I developed the suspicion that two other of his works —Angela the Orphan, or The Bandit Monk of Italy and The Black Forest, or The Solitary of the Hut— must also be reprints. The British Library copy of the first having been digitized, I was able to search for passages in the Corvey Collection, which is part of Nineteenth Century Collections Online published by Gale, and found that it was in fact The Italian Marauders. Spotting a pattern had already led to new findings.

The only extent copy of The Black Forest is held at the University of California, Los Angeles. I received from the library staff photographs of the first pages in the spring of 2021 but have not been able to find a source text yet in digital archives. The first two steps were taken in order to produce an abstract to submit for the collection.

Step 3. Identifying other 19^th-century reprints of these titles

To understand if the choices of penny publishers were reproducing previous reprinting patterns and/or setting new ones, I set out to reconstruct the publishing history of each of the novels identified in Step 1. I moved from the ODS spreadsheet to a Google Spreadsheet.

I first consulted Montague Summers’s Gothic Bibliography, adding each new publisher or series as a column. I included the volume number or the year in which the reprint was issued in the cell at the intersection of the row of the title and the column of the publisher or series. If neither was provided, I put an “x”. I supplemented this initial research with other bibliographic work by Alice M. Killen, Deborah D. Rogers on Ann Radcliffe, Gary Kelly on Clara Reeve, and William St Clair on William Godwin.

In the end, my spreadsheet contained 48 columns dedicated to series, 35 for the first half of the century (including four for penny weekly reprints) and 13 for the second half of the century.

Step 4. Checking for other potentially Gothic titles reprinted alongside

As I was researching Milner’s series, I noticed Gary Kelly included in his Varieties of Female Gothic Jane Porter’s Scottish Chiefs, which I recognized from my database, but which Hannah-Freya had not flagged as Gothic. It had also been reprinted in Bentley’s Standard Novels along with another novel of hers as well as The Hungarian Brothers by her sister Anna Maria which I recognized from the Novel Newspaper. Excluding these three novels meant it seemed like Bentley’s had not reprinted much first-wave Gothic content.

I ended up adding the contents of the first 11 volumes of Bentley’s series (with the exception of the four American titles and another by Friedrich von Schiller) and Vathek, the third novel in volume 41. Since Anna Maria Porter was brought into the fold, I also added a second novel of hers that Bentley’s had not reprinted, but which appeared in the Novel Newspaper. This operation added five rows to the spreadsheet, including Mary Shelley’s Frankenstein, which had not been reprinted in penny weekly numbers, probably because the penny publishers feared litigation from Richard Bentley.

The spreadsheet in this state forms dataset B.

Step 5. Selecting the most important 19^th-century publishers and series in first-wave Gothic reprinting

Since there were too many series (48!) to all discuss in our book chapter, I selected only the series containing the most Gothic titles. I tallied up the number of rows in each column which contained a value, be it a volume number, a year, or an “x”. Below is the distribution of the number of reprinted titles in each series.

On the left-hand side, we see many series (12) contained a single title from the list of 35 potentially Gothic novels. Hannah-Freya and I would only talk about those with the most titles, on the right-hand side.

The question was how far left to go, where to draw the line between important enough by virtue of reprinting a high number of first-wave Gothic novels, and not important enough to be discussed. There are five series with five titles, only two contain six, a single one seven titles, whereas there are three which contain eight reprinted titles, which I chose as the cut-off.

This culling process would have reduced the dataset to 11 series/columns. However, I kept Barbauld’s British Novelists given that it served as a basis for Walter Scott’s Ballantyne’s Novelist’s Library, as well as T. Paine and John Lofts because they published in penny weekly numbers, the subject of our chapter. I eventually removed Warne’s Notable Novels, which contained eight reprinted novels, but three of these were by the Porter sisters. In the end, dataset A includes 13 series/columns.

Step 6. Removing less pertinent novels

As the Gothic scholar in the team, Hannah-Freya was tasked with investigating the new additions from Step 4. Reviewing recent scholarship on the Porter sisters’ novels, she came to the conclusion that they should not be included in the dataset. The four of them were therefore removed from the final product, dataset A.

Dataset B contains two other novels which were removed in dataset A: Vathek —which was reprinted in volume 41 of Bentley’s Standard Novels, but in only two other editions, both in the second half of the 19^th century— and Thalaba the Destroyer, which had served as the source of an adaptation in penny weekly numbers rather than a reprint.

Step 7. Grouping the novels into categories

A failed attempt at creating a Sankey diagram showing how titles moved from one series to the next led me to create categories for the novels.

1. I dubbed the most reprinted novels throughout the 19^th century, often reprinted together —The Old English Baron and The Castle of Otranto— the Founding pair.
2. Ann Radcliffe‘s five novels formed their own group.
3. Unable to find other novels with which to group it, I placed Charlotte Smith’s The Old Manor House on its own.
4. I called the group of four novels that took over Radcliffe’s in popularity in Milner’s series the Later 19^th-century Gothic bestsellers.
5. Other Gothic novels reprinted by Newman is the largest group with an heterogenous subset of nine novels that did not appear in Barbauld’s collection nor Ballantyne’s nor Bentley’s series, but had been reprinted by Anthony King Newman, successor to William Lane at the Minerva Press.
6. Bentley’s provided a better defined category, containing the four novels reprinted in his Standard Novels that had not appeared in a more popular category, i.e. The Castle of Otranto which is in the founding pair. It could also have been called “The Godwins” —for William Godwin and Mary Shelley, née Godwin— if it were not for the presence of Matthew Lewis’s translation The Bravo of Venice.
7. The Monk was placed in its own category given its centrality in Gothic scholarship but its relatively low degree of reprinting in the 19^th century.
8. The last category contains three novels for which the first and only reprints were published in penny weekly numbers (no reprints under Newman’s imprint could be located in Worldcat). They all happen to have been originally published by J. F. or George Hughes.

Milner’s series

In an attempt to trace the evolution of first-wave Gothic reprinting in the second half of the 19^th century, represented only by the Milner’s series column in dataset A, I created a new spreadsheet to separate out the contents of the various series published by the firm.

Dataset C is thus the result of searching for the titles in Dataset B in two different digitized catalogues of the Milner publishing firm:

• Cheap Books Published by Milner & Sowerby, Paternoster Row, London, and Sold by All Booksellers (London: Milner & Sowerby, between 1866 and 1874), 32 p. Google Books.
• A Retail Catalogue of Books (London: Milner & Co., 1893 or 1894), 32 p. Google Books.

The range of possible dates for the first catalogue is established by comparing its imprint to the research contained in William Milner of Halifax, printer and publisher, obtained with the help of Claude J Pelletier who blogged about his quest to learn more about Milner. The second catalogue, also undated, was bound with other publishers’ lists dated 1893 or 1894.

Both catalogues contain multiple series or “libraries”. Each title was therefore searched for in each series of the two catalogues.

Dataset description: Presenting the data

All three datasets contain novels listed in rows and collection, publishers, or series listed in columns. If the novel in a given row was reprinted in the collection or series or by the publisher in a given column, the cell where they intersect will contain a date (the earliest reprint of that novel found in that collection or series or by that publisher), a volume number, some other information, or simply an “x”.

Dataset A

Dataset A contains 29 novels listed in rows and contained in a total of 13 collection, publishers, or series listed in columns. The novels are grouped into eight categories, with subtotal rows preceding the list of novels in each category.

The last two rows give an overall view. The penultimate row contains the total of first-wave Gothic novels in each series and the last row gives the share of all 29 novels reprinted in the series as a percentage.

Novels

The 29 novels are identified in the first three columns by their:

1. author;
2. title;
3. initial publication year.

They are organized into categories and, within them, sorted in chronological order of first publication. The groupings are described in Step 7 above. In the formatted spreadsheet, Radcliffe’s novels are shaded in the darker grey and the group of later 19^th-century Gothic bestsellers in the light grey, as the novels individually are in Dataset B.

Series

The 13 series are identified in the first three rows. They provide:

1. Its title or the name of the publisher or the editor;
2. The year in which the first volume was published;
3. The price point, if regular.

The series are organized chronologically. In the formatted spreadsheet, the Victorian series are coloured in red.

Totals

The last two rows give an overall view. The penultimate row contains the total of first-wave Gothic novels in each series and the last row gives the share of all 29 novels reprinted in the series as a percentage.

Dataset B

Dataset B contains 35 potentially first-wave Gothic novels listed in rows and reprinted in a total of 48 publishers, collections, or series listed in columns. The series are organized chronologically and split into the first half of the century (35) and its second half (13).

The last column indicates the total number of series in which a novel appears with columns indicating subtotals for each half-century. The last row shows the total number of potentially first-wave Gothic novels reprinted in each series.

Novels

The 35 novels are identified in the first seven columns by their:

1. author;
2. title;
3. initial publication year;
4. initial publisher;
5. ESTC number;
6. DBF record number at British Fiction 1800-1829;
7. identifier in Corvey Women Writers on the Web (CW³), no longer accessible.

They are sorted from most reprinted to least reprinted with a preference for those most reprinted in the first half of the century. Novels that are present in the same number of series overall and in the first half of the century are ordered chronologically.

Radcliffe’s five novels are shaded in the darker grey and the four later 19^th-century Gothic bestsellers in the light grey.

Series

The 48 series are identified in the first three rows. They provide:

1. Its title or the name of the publisher or the editor;
2. The last two numbers of the year in which the first volume was published;
3. The price point, if regular.

The series are organized chronologically. In the formatted spreadsheet, the most important series’ names are in bold and, of those, the Victorian ones are coloured in red, like in Dataset A.

Totals

The last column indicates the total number of series in which a novel appears with columns indicating subtotals for each half-century. The last row shows the total number of potentially first-wave Gothic novels reprinted in each series.

Dataset C

Dataset C contains 15 novels listed in rows and contained in a total of 12 “libraries” or headers from two catalogues of the Milner publishing firm, listed in columns. The last column indicates the total number of series in which a novel appears while the last row shows the total number of first-wave Gothic novels (excluding those written by the Porter sisters) reprinted in each series.

Novels

The 15 novels are identified in the first two columns by their:

1. author;
2. title;
3. initial publication year.

They are sorted from those present in the most series to those in the least series.

The three novels by the Porter sisters are struck through in the formatted spreadsheet. The series in which they are included are indicated with an “x” rather than a “1” so as to not effect the count of first-wave Gothic novels in each series.

Series

The 12 series are identified in the first four rows. The first two rows separate the columns according to the catalogue in which the series is found. They give:

1. The hyperlink to catalogue digitization in Google Books;
2. The date range for the possible date at which the catalogue was issued.

The next two rows are series-specific. They provide:

3. The title of the series as indicated in the catalogue;
4. The page range in which the contents of the series are listed in the catalogue.

Within each catalogue, the series are ordered from the one containing the most first-wave Gothic titles to the least.

Totals

The last column indicates the total number of series in which a novel appears. As indicated above, the last row shows the total number of first-wave Gothic novels reprinted in each series excluding those written by the Porter sisters.

Data illustration: Representing the data

Sorting through spreadsheets can only help you so much: to make sense of data, I need to visualize it. Between steps 3 and 4, wanting to understand titles according to how they were grouped by the 19^th-century publishers themselves, I hand-drew Venn diagrams with the titles as points organized into subsets representing the series. I realized Walter Scott was drawing from Anna Læticia Barbauld’s selection, forming an early 19^th-century first-wave Gothic canon John Limbird also built on.

I knew preparing nice diagrams would be time-consuming, so for the book chapter I first attempted to present the data in a series of four tables each combining two series. The UWP style guide however made this solution very impractical, so I returned to the idea of representing it in diagrams, now knowing the arguments we were making in the chapter.

I started again by hand-drawing my diagrams. Because they are not produced by an algorithm, Stephen Webb suggests I call them data illustrations rather than data visualizations. The first figure shows the general movement away from Radcliffe when comparing the early 19^th-century first-wave Gothic canon (which I called “Romantic”, but Hannah-Freya pointed out can cause confusion) to that in the later half of the century (which I called Victorian), incorporating a set of now non-canonical works, the “Later bestsellers” at the bottom of the data illustration.

Because of the issues with naming these two versions of the canon, we voted against incorporating it into the chapter. Above is a scan of my hand-drawn data illustration.

The next data illustration —placing on timelines the novels reprinted in two late 1830s cheap series— made it into the chapter, so I will not comment on it. Stay tuned for its forthcoming publication!

The last data illustration —representing the reprints in penny-weekly number-books— did not make the cut. It groups titles into buckets, ordering them from most reprinted overall in the 19^th century at the top to least reprinted at the bottom.

On the right-hand side, in the large shaded area, are the six titles reprinted by “T. Paine”. Four of these represent to whole subset known in dataset A as “Later 19^th-century Gothic bestsellers”. The other two are part of the last group in dataset A, known as “Novels published by Hughes”. All novels initially published by J.F. or George Hughes are shaded in darker grey.

On the left-hand side are all the first-wave Gothic novels that T. Paine did not reprint, but that other publishers released in penny-weekly numbers. Three had previously appeared in Barbauld’s British Novelists.

Lewis’s highly influential The Monk is reprinted in two penny-weekly editions under its original title, a novelty since chapbook adaptations and cheap reprints had changed the title to reflect specific subplots, like The Bleeding Nun and Raymond and Agnes. Other publishers still reprinted Lewis’s translation, The Bravo of Venice, which previously appeared in volume 41 of Bentley’s Standard Novels, and Emmeline, reprinted in the Novel Newspaper.

Finally, penny publishers innovated by reprinting first-wave Gothic novels that had either been reprinting by Newman, either not reprinted at all since their first publication. Of these seven titles, more than half were initially published by J. F. or George Hughes, who was not considered respectable in the 1810s even though he occupied a large share of the market for fiction.

Figure 3 shows how hard it can be to neatly arrange some slices of data.

---

---