The FAQ Is Not the Rule

There is a page on the Texas Department of State Health Services site called Consumable Hemp Products — Frequently Asked Questions. Thirty-one questions. It is the page you land on when you type an ordinary operator question into a search box, and several of the answers we checked this week cite it.

We put its URL in our own research notes on 21 August, with a summary we had taken from somebody else's write-up of it. Then we did not open it for thirty-seven days.

We opened it today and read all thirty-one answers. Two of them are worded differently from the rule they are answering about, and one of them leaves out the provision sitting immediately beside it.

First, the thing that makes this hard to notice

The page carries no date.

Not a "last updated". Not a "reviewed on". Nothing in the markup either — we checked for a publication or modification date in the page source and there is none.

So the page contains sentences about a rule change that took effect in March 2026, and sentences about a court decision from 2022, and a sentence with an effective date in March 2023, all in the same typeface with the same authority and no way to tell which was written when.

We wrote about the limits of date stamps a few days ago, at gethempdash.com/learn/a-date-stamp-is-not-a-warranty. That piece made the case that a date stamp tells you when something was written and not whether it is still true. This is the harder version of the same problem and we are not going to re-argue it: here there is no stamp to distrust.

Divergence one: which THC number

Here is what the FAQ says, quoted, in answer to "What are the requirements to sell CHPs in retail?"

> CHPs sold must not contain more than 0.3 percent of Delta-9 tetrahydrocannabinol (THC).

And in answer to a question about out-of-state manufacturers, it says a firm should have a certificate of analysis that "proves the CHP does not exceed 0.3 percent Delta-9 THC content."

Now here is the definition from the adopted rule, 25 TAC 300.101(1), quoted in full:

> Acceptable hemp THC level--A total delta-9 tetrahydrocannabinol content concentration level on a dry weight basis, that, when reported with the accredited laboratory's measurement of uncertainty, produces a distribution or range that includes a result of 0.3% or less.

The defined term in the rule is total delta-9. The same chapter separately defines "Delta-9 tetrahydrocannabinol (d-9 THC)" as the isomer — a different defined term, at a different number, in the same definitions section.

The word "total" does not appear anywhere on the FAQ page. We counted: zero occurrences across all thirty-one answers.

This matters because the two are not the same measurement. Total delta-9 THC is a computed value — it adds a share of the THCA on the certificate to the delta-9 figure. A result that sits comfortably under the line on the delta-9 row alone can land somewhere else entirely once the computation is done. The arithmetic is at gethempdash.com/learn/total-delta-9-thc-how-it-is-computed and we are not restating it here, because nothing about it has changed and that page does the job.

What we will say is the narrow thing. If you take the FAQ's sentence as the test your certificate has to pass, you are reading one row where the rule's defined term is built from two.

We are not saying the FAQ is wrong. We do not know which instrument its sentence is tracking — the hemp definition in the Texas Agriculture Code is written around delta-9, and the rule's product threshold is written around total delta-9, and the page does not say which one it means. That ambiguity is the finding. We are not going to resolve it for you, and anyone who resolves it confidently in either direction is doing something we could not do with both documents open.

One more thing worth recording, because we had been getting this wrong about ourselves. For three mornings we logged the missing word as a defect in machine-generated answers — somebody's summariser dropping a qualifier. It is on the agency's own page. We are not claiming those summaries took it from there; we cannot show that and will not assert it. But we had been describing a first-party wording problem as a third-party one.

Divergence two: the age check

The FAQ:

> A person who sells a CHP must verify each purchaser is 21 years of age or older by inspecting a valid proof of government-issued identification before completing the sale of any CHP.

The rule, 25 TAC 300.701(b):

> A person who sells CHP must verify each purchaser's age by reviewing a valid proof of identification before completing the sale of any CHP.

Read them next to each other. The rule says verify the purchaser's age. The FAQ says verify the purchaser is 21 or older. Those land in the same place, because the chapter defines a minor as a person under 21 years of age, so we are not going to make more of that difference than it deserves.

The detail we want to flag is elsewhere. The FAQ's age answer cites no section at all. On the same page, the packaging answer cites 25 TAC 300.405, subsections (1) through (3), by number. The answer about destroying non-compliant product cites 25 TAC 300.604 and 300.605 by number. The age answer — the one that governs what happens at the counter on every single sale — cites nothing.

That is what sent us looking for where the age duty actually lives, and the search turned up something worth passing on.

The provision the FAQ leaves out

The age requirement sits in a subchapter of its own — Subchapter G, "Restrictions on Sale to Minors", new sections 300.701 and 300.702, adopted under TRD-202601066 and published in the Texas Register of 20 March 2026 with an effective date of 31 March 2026.

Section 300.701 sets out the duty, including what counts as valid identification: a driver's licence from Texas or another state, a passport, or a state or government agency identification card, and it must carry a photograph matching the person, show a date of birth, be issued by a government agency, and not be expired.

Then there is 300.702, which the FAQ does not mention. Quoted:

> (a) The department may, after providing an opportunity for a hearing, revoke a consumable hemp license or retail hemp registration after determining the license or registration holder, or an employee, sold, served, or delivered a consumable hemp product to a minor. (b) An exception to subsection (a) of this section exists where the minor falsely represents to be at least 21 years of age by displaying an apparently valid proof of identification.

Subsection (b) is a provision about what happens when a seller checked an identification and the identification was a lie.

We are not going to tell you what that provision would do for you, and we want to be plain about why. How it applies to a particular sale, what "apparently valid" requires, what a shop would need to be able to show — those are questions for a lawyer looking at your facts, not for us. What we can tell you is that the provision exists, that it sits in the same subchapter as the duty, and that it is not on the page you are most likely to read first.

That is the point of this article in one line. The FAQ answers the question. The subchapter contains the answer and the thing next to the answer.

Where the section numbers do appear

A note for anyone who goes looking, because the trail is misleading.

The only DSHS page we found that names 300.701 and 300.702 by number is a news release dated 3 October 2025. It announces those sections as emergency rules, "effective immediately", and states in its own words that they "may remain in effect for up to 120 days with the possibility of one 60-day extension."

Count forward from October 2025 and that instrument ran out in the spring.

The permanent rules did follow — proposed in December 2025, adopted effective 31 March 2026 — and they are the ones we quoted above. So the duty is live under the permanent rule. But notice the shape of what a searcher walks into: the page that dates itself points at an expired instrument, and the page that states the current duty carries no date and no citation. Neither page is hiding anything. Between them they make it genuinely difficult to answer "which rule, as of when".

A near-miss we should record

Verifying 300.702 took three documents instead of one, and there is a lesson in why.

Section 300.701 was adopted with changes, which means the Texas Register republished its text. Section 300.702 was adopted without changes — and the adoption says, in as many words, that such sections "will not be republished." So the operative text of 300.702 is not in the adoption at all. We verified it against the December 2025 proposal, which the adoption's own "without changes" statement makes good.

If you only ever open adoptions, you cannot read half of this subchapter. That is not a quirk of this rule; it is how the Register works.

And the near-miss. DSHS publishes a consolidated chapter text as a PDF, and we read it. It starts at 300.100 and runs to 300.702, and it looks complete. Section 300.104 is not in it. We were close to writing that down as a finding — until the adoption's own comment responses turned out to reference "25 TAC 300.101-300.104". The section exists. The PDF is a partial consolidation that begins at the beginning, which is exactly what a complete document also does.

The practical version

Ask which surface a requirement came from, not just which agency. "DSHS says" is not one answer. There is a programme page, a schedules page, a frequently-asked-questions page and a news-release archive, and this week they were not saying the same thing in the same words. We wrote a piece on asking which rule someone means, at gethempdash.com/learn/under-which-rule. The case it does not cover is this one, where both answers come from the same agency.

When guidance paraphrases a rule, read the subchapter the rule sits in. Not the cited section — the subchapter. The duty and the provision beside it are frequently not in the same sentence, and a paraphrase has to choose.

A defined term is worth more than a plain-English restatement of it. "Acceptable hemp THC level" is defined once, in one place, and it is built out of a computation. Every sentence in the chapter that turns on that phrase inherits the computation. A restatement that swaps in the ordinary-language version drops it silently.

If the page carries no date, the only date on it is the one you write down. We read this FAQ on 27 September 2026. That sentence is the whole provenance for everything we have quoted from it, and it is why we put it in writing.