EML Exhibit
Guide

How to prepare email exhibits for trial

A production is built for opposing counsel and a review platform. An exhibit is built for a witness, a judge and a jury, and it is read once, under time pressure, by someone who has never seen the thread before.

Those are different jobs, and the second one is mostly about editing: deciding what a single exhibit is, what it must show on its face, and what can be left out without giving the other side a completeness objection.

Last reviewed
Not legal advice. This page describes common practice and general procedure. Rules of evidence, local rules and their interpretation vary by jurisdiction and change over time; confirm the requirements that govern your matter before relying on any of it.

An exhibit is an editing problem

By the time email reaches this stage the technical work is done. The messages were collected, converted, numbered and produced. What remains is a judgement about presentation, and it is made against a specific reader: someone encountering this conversation for the first time, in a courtroom, while also tracking the question being asked.

That reader has no context, cannot scroll, and will not re-read. Everything below follows from that.

What one exhibit should be

The first decision, and the one that shapes everything else.

A single message works when the examination turns on a specific statement — what the witness wrote on a particular day. The exhibit is short, the witness cannot wander into surrounding traffic, and the point is immediate.

A whole thread works when the sequence is the evidence: who was told what, in what order, and what happened next. This is the more common shape in practice, because email evidence is usually about knowledge and timing rather than about a single sentence.

Pick a convention and hold to it across the case. An exhibit list where some entries are messages and some are threads, apparently at random, is harder for everyone — including you, at 9pm the night before.

Produce it from the native

Render exhibits from the retained .eml files rather than from a screenshot, a forwarded copy, or a printout someone made months ago.

The reason is that the render puts the header block on the face of the document, in the standard form:

From:     Alison Reyes <a.reyes@example.com>
To:       David Whitlock <d.whitlock@example.com>
CC:       Priya Nandakumar <p.nandakumar@example.com>
Date:     14 March 2025, 09:12
Subject:  Re: Q1 forecast — revised numbers

An exhibit carrying that block describes itself. Handed to a witness, it answers who, to whom, and when without anyone testifying to it. An exhibit showing only body text needs a witness to establish what it is, which is fine until the witness is unavailable, evasive, or genuinely cannot remember.

This is also why forwards make poor exhibits: the header block on a forward is the forward’s, and the message you actually care about is quoted inside the body under someone else’s envelope.

Order the thread forwards

Mail clients display conversations newest-first, because that is what is useful when you are working through a mailbox. It is exactly wrong for an exhibit.

A reader meeting the conversation for the first time needs it oldest-first, so that it reads as events in sequence. Reversed, the reader has to reconstruct the chronology while also following the examination, and the effect of the sequence — which is usually the point — is lost.

Conversion tooling that can sort by the date on the message rather than by filename does this in one step. See converting EML to PDF.

Keep the Bates numbers

Trial exhibits carry their own numbering — Exhibit 14, Plaintiff’s Exhibit C, whatever the convention is in that court. The underlying Bates numbers from the production should stay visible on the page as well.

The reason is practical. During examination someone will need to establish that this exhibit is a document that was produced, and doing that by pointing at a number on the page is instant. Doing it by comparing content is not.

This also means exhibits should be cut from the produced set rather than re-rendered from the natives at exhibit-preparation time. A re-render can paginate differently, and an exhibit whose page count differs from the produced version is a question you do not want asked.

Attachments

If the attachment matters, it has to be present. An exhibit consisting of an email saying “see attached” with no attachment is a gap a witness will find.

Two workable conventions: the attachment as part of the same exhibit, immediately following the covering message; or the attachment as its own exhibit, cross-referenced. Either is fine. What matters is that the attachment is the original file as produced, not a re-rendered copy — the version the recipient actually received. See parent-child email productions.

Verify against the production

The last step before the exhibit list is fixed: check each exhibit against the produced version, page for page.

What you are looking for is any discrepancy between what opposing counsel received and what you are about to put in front of a witness — a different page count, a missing quoted section, a different rendering. These arise innocently, usually from an exhibit prepared from a native at a different time with a different tool, and they are unpleasant to have raised mid-examination.

What stays in the file

The natives. Always, and past the end of trial.

If authenticity is genuinely contested — a witness denies sending a message — the argument moves to the transport headers, and those exist only in the retained .eml. The exhibit shows what the message said; the native shows how it arrived. Email metadata and admissibility covers what that argument looks like.

This is not legal advice. Exhibit practice, authentication requirements and completeness objections vary by jurisdiction and by the rules and standing orders that govern your matter. Confirm before relying on any general description.

Doing it in EML Exhibit

  1. Decide what one exhibit is

    A single message, or a whole thread. Single messages for pinpoint impeachment on a specific statement; whole threads where the sequence is the point. Mixing the two conventions across a case makes the exhibit list hard to follow.

  2. Produce it from the native, not from a printout

    Render the exhibit from the retained .eml so the header block appears on its face in the standard form. An exhibit that shows sender, recipients, date and subject is self-describing.

  3. Keep the Bates numbers visible

    Trial exhibits carry their own numbering, but the underlying Bates numbers should remain on the page so the exhibit can be tied back to the production without argument.

  4. Order a thread chronologically

    Mail clients display newest first. A reader meeting the conversation for the first time needs oldest first, so the exhibit reads as a sequence of events rather than backwards.

  5. Check it against the produced version

    Before it goes on the exhibit list, confirm the exhibit matches what was produced, page for page. A discrepancy between exhibit and production is an avoidable and damaging thing to have raised mid-examination.

Questions
Should a whole thread be one exhibit or several?

It depends on what you are doing with it. If the examination turns on a single statement, a single-message exhibit keeps the witness on that statement. If the point is what someone knew and when, the thread is the exhibit, because the sequence is the evidence. What causes trouble is switching conventions arbitrarily within one case.

Can we exclude irrelevant messages from a thread exhibit?

Cautiously. Trimming a thread invites a completeness objection and the suggestion that context was removed. Where a thread contains genuinely unrelated traffic, it is usually cleaner to use the full thread and direct the witness to the relevant message than to produce an edited version you then have to defend.

Do the quoted portions need to be included?

Usually yes, and they come along automatically since they are part of the message. The awkward case is a long quoted tail that repeats material already in the exhibit. It is repetitive rather than improper, and removing it is an edit you would have to justify.

What about attachments?

If the attachment matters, it is part of the exhibit or is its own exhibit, and it should be the original file rather than a re-rendered copy. An exhibit consisting of a cover email referring to an attachment nobody has is a predictable problem during examination.

Does the exhibit need the full email headers?

The face-of-the-message block — from, to, cc, date, subject — should be visible. Full transport headers are not normally put in front of a jury; they live in the retained native and come out only if authenticity is genuinely contested. See email metadata and admissibility.

EML Exhibit

Run this workflow without the manual steps

EML Exhibit does the render-and-batch part of the process above — headers printed above the body, attachments preserved byte-for-byte, output ready for the stamp.

Try it on a batch
Related