The Audley End Annotations

What the Evidence Shows

Audley End House in Essex was built in 1605–14 by Thomas Howard, 1st Earl of Suffolk. From 1797 it belonged to the Barons Braybrooke, and in 1948 it was sold to the Ministry of Works; English Heritage now manages it, and most of the books still belong to the Braybrooke family. Many of the older books, including those discussed here, are said to have come to Audley End from Billingbear in Berkshire, a Neville family house, after it burned down around 1924.

Seven sixteenth-century books in the collection (six to eight, depending on how bound volumes are counted) contain more than 1,500 annotations by the 2023 paper’s count (Stritmatter’s 2025 essay puts it at over 1,800), mostly in Latin or Greek, a few in Italian or English. Examination has identified three distinct hands: Sir Henry Savile (1549–1622), Sir Henry Neville (1564–1615), whose hand appears in only one volume, and a third annotator previously labelled the “Audley End Unknown.”

Roger Stritmatter, Professor of Humanities at Coppin State University, published a 2023 paper in the Journal of Forensic Document Examination (an 18,000-word study applying Huber and Headrick’s eight elements of handwriting discrimination, plus two further standards: whole-word comparison and, for the positive case only, pen control). It proceeds in two stages: it first tests Neville as the Unknown annotator and rejects him (against the earlier hypothesis of Casson and Rubinstein), then tests de Vere, and concludes that the annotations were “with very high probability” written by him (the abstract puts it more cautiously: “de Vere is a more plausible writer of the Audley End annotations”). The journal is an internationally peer-reviewed publication with no connection to the authorship debate. Stritmatter’s 2023 study is the most methodologically rigorous forensic work in the Shakespeare Authorship Question. It is a single examiner’s finding: the paper asks other document examiners to test it independently, and none has yet reported. On that footing, and whatever one concludes about authorship, it is a real addition to the historical record.

The annotations include Latin glosses on difficult Greek words and corrections to the Greek text of the Estienne editions, which require the ability to read printed ancient Greek closely. The Cassius Dio, the most consequential volume, is the Stephanus edition of 1548, physically bound with the 1551 first edition of Appian’s history of Rome (whose notes are Savile’s). The same hand also appears in a 1574 Tacitus. The 2023 JFDE paper addresses whose hand made these annotations. Stritmatter proposes correspondences between the annotated sections of Cassius Dio and dramatic material in Julius Caesar and Antony and Cleopatra; Dennis McCarthy has contested them in a detailed online rebuttal (Substack, August 2026), arguing that the annotator shows no particular focus on the events the plays dramatise and that where Dio and North differ, the plays follow North. A follow-up specifically surveying the literary correspondences between the annotations and the Shakespeare plays has not appeared in peer-reviewed form; as of 2026 it exists only in a 2024 online essay and a 2025 Substack series, which McCarthy’s rebuttal has since contested.

The forensic method is suited to what it tests: handwriting attribution. The questions are how firm the attribution is, and what it would prove about authorship if it holds.

The Inferential Chain

The argument from annotation to authorship requires three steps.

Step 1: The annotations are Oxford’s hand. This is the step the forensic study addresses, and it is the strongest part of the argument. The study compares the Unknown hand with letters by Neville and by de Vere, checks how rare the shared features are across more than 260 sixteenth-century hands and a similar number of copybook models, and concludes that de Vere is the writer “with very high probability”. It tests the one alternative proposed in print, Neville, rather than a wider set of named contemporaries, and no independent replication has yet been published, though the paper invites one. It applies standard forensic handwriting criteria in a specialist peer-reviewed journal; a reader who wishes to challenge Step 1 must engage it on its own forensic grounds. A reader who would accept an attribution to Neville by these standards is obliged to weigh one to de Vere by the same standards. The examination is also separate from Stritmatter’s earlier study of de Vere’s Geneva Bible, and an objection to that study does not carry over to this one. This site cannot adjudicate the handwriting evidence itself; what follows concerns what the attribution would prove if it holds.

Step 2: The annotated passages correspond to the plays. Stritmatter argues that the Cassius Dio sections the annotator marked are the ones that correspond to the dramatic action of Julius Caesar and Antony and Cleopatra; McCarthy’s rebuttal disputes that the annotator’s choices track the plays at all. If the correspondence holds and the hand is Oxford’s, it shows he read and marked those passages. Who wrote the plays is a separate question.

Step 3: Therefore Oxford wrote the plays. This step is not supported by the evidence alone. It requires an additional premise, that the person who annotated the source was also the person who turned it into drama, which is precisely what the whole debate is about. The discovery is consistent with the Oxfordian hypothesis. It is also consistent with Oxford being a learned reader whose books were accessible to another writer, a patron who discussed sources with a writer in his circle, or any number of other arrangements. Consistency with a hypothesis is not proof of it.

The Specificity of the Correspondence

The Oxfordian case here goes beyond Oxford having read Cassius Dio: Stritmatter argues that the annotator marked the specific passages corresponding to the dramatic action of Julius Caesar and Antony and Cleopatra rather than annotating the text uniformly. That selectivity, if confirmed, would be a stronger observation than the simple fact of having read the source.

Four things bear on it.

First, the full literary analysis has not yet been published. The 2023 JFDE paper addresses handwriting attribution. The follow-up on the literary correspondences has not appeared in peer-reviewed form: as of 2026 Stritmatter has set the correspondences out only in a 2024 online essay and a 2025 Substack series, and McCarthy’s rebuttal has since contested it. The precise scope of the correspondence (how many annotations, what proportion of the full Cassius Dio text, and how closely they track what became dramatic scenes) has not yet had a full comparative analysis in peer-reviewed form, though readers can set the annotated passages against Dio, North and the plays. This site will update when it does.

Second, even granting that the correspondence is highly specific, it is consistent with two explanations. An author marks the passages he intends to dramatise. A patron marks the passages he wants his company’s playwright to dramatise. Either could produce annotation of this kind. What might tell them apart (the density, choice, order and wording of the notes) has not yet been analysed in any peer-reviewed study. The selection would establish purposeful engagement with those passages and leave open whose purpose it was.

Third, the Greek corrections add to the picture without settling the question. Corrections to a printed classical text in the Estienne editions require real fluency in ancient Greek, so, if the hand is Oxford’s, they strengthen the evidence of his close engagement with the material. They do not show whether he wrote from it or directed someone else to.

Fourth, the specificity of the correspondence has to be weighed against the plays’ established source. The documented primary source for both Julius Caesar and Antony and Cleopatra is Plutarch (specifically North’s 1579 English translation, whose verbal influence on the plays’ dialogue is direct and traceable line by line). Enobarbus’s description of Cleopatra’s barge (II.ii) follows North’s prose almost phrase for phrase. Where Dio and North tell the same events, a correspondence counts as evidence that the plays drew on Dio only where they take a detail from Dio that North does not supply.

The Patron Argument

The “(s)he could have” rule stops people using possibility to build a case. It has nothing to say when documented practice is used to limit what someone else’s evidence proves. The Oxfordian inference would be that the annotation establishes authorship. This site’s response builds no case for anyone, so it falls outside the (s)he could have rule: it shows that the annotation is consistent with authorship without establishing it, because patron-directed writing is documented Elizabethan practice. The patron route is offered only as a possibility that stops the annotation being decisive. No document connects Oxford, these books and a playwright, and this site does not put it forward as what happened.

Three documented instances establish that practice.

First, for Elizabeth I’s visit to Kenilworth in 1575, Robert Dudley, Earl of Leicester, had speeches and verses devised for particular moments of the visit by George Gascoigne and others; they were printed that year as The Princely Pleasures at the Court at Kenilworth.

Second, Philip Sidney states in his own dedication that his sister “desired” him to write the Arcadia.

Third, Samuel Daniel’s dedication to Cleopatra (1594) explicitly states that the Countess of Pembroke “did impose” the work on him.

Three documented instances, across different patrons, writers, and genres, of the same underlying practice: patrons directing writers at the level of specific content. None of them involves an annotated source text, so the specific idea that a patron marked classical sources for a dramatist’s use has no documented instance. The examples show only that patron direction was documented, which is enough to stop the annotation counting as decisive.

A possible response (if Oxford was directing a playwright, why did the book stay with Oxford rather than passing to the writer?) would be less decisive than it appears. Elite libraries in the period were routinely used by writers and scholars without the books leaving the owner’s possession. Sir Robert Cotton’s library is a well-documented case: leading scholars, Bacon and Ralegh among them, used it, and the library remained Cotton’s. How these books came to be at Audley End is itself part of the evidentiary record Stritmatter addresses. Their location tells us about ownership and descent, and nothing about who made intellectual use of them.

The patron-supply hypothesis is itself inference: no primary source records Oxford directing a playwright to these passages, or any writer making use of his annotated books. Set out plainly: if the hand is Oxford’s, the annotation establishes that he read and marked this history closely. It makes more probable that he had the learning and the books to engage with the plays’ Roman material, which the authorship case needs, and if the correspondences Stritmatter proposes hold, it would also show deliberate engagement with the passages the plays use, which is more specific. What remains merely possible, on present evidence, is either route from his reading to the plays: that he wrote them, or that he supplied a writer. The surviving record documents neither, and this site does not rank the patron route above the authorship one.

What Would Close the Gap

The gap is between Oxford the reader and Oxford the author of the plays. An annotation in a source text connects a person to their reading, not to their writing. What would close it is a document in Oxford’s hand connecting him to the Shakespeare plays themselves (a draft passage, a letter mentioning the works) or a contemporary who credits him with one of the plays by name.

How the annotations score in Diana Price’s framework is a separate matter. They fall under Category 5 (Original manuscript extant), where they are the only entry on this site that rests on a forensic attribution, and they carry the F.A. (Forensically Attributed) designation, scored at the same weight as a partial (◑) finding. Category 5 asks whether any original literary manuscript by the candidate survives, of any of his own works. The annotations are in his hand, on the forensic record, but they are notes on a work of history, so they meet the handwriting half of the test and not the literary half. A poem in de Vere’s own hand would lift the score to ✓, and it would do nothing to close the authorship gap: that he wrote poetry is already documented, and the Paper Trail measures whether a candidate left the traces of a working writer, not whether he wrote Shakespeare.

If this site calls the Audley End work rigorous, significant and a real addition to the record, what result would ever be enough? A document in Oxford’s hand connecting him to the composition of the plays. No further forensic examination of Oxford’s reading, however rigorous and however many more annotated volumes are found, can supply direct evidence of him writing them, although a large, well-matched body of such annotation could add materially to the circumstantial case. Short of such a document, evidence like this still counts: the site’s convergence analysis weighs it as circumstantial. What it cannot do is settle the question: that needs a record of the writing, or circumstantial evidence that meets the five conditions on What Would Settle It, and annotations in books de Vere read would fit other readers of those books too. That is why the site can praise the study and still decline to treat it as proof of authorship.

Full citations for this page are on the Sources page.