If your page ranks on the first page of Google and still does not appear when someone asks ChatGPT or Gemini the same question, you are not imagining the gap. The gap is real, it is documented, and it has a specific cause that is almost never the one being sold to you.
The reason most explanations miss is that they treat ranking and citation as the same problem with different dials. They are not. They are two separate eligibility tests, run by different systems, against different criteria. Passing the first says nothing about passing the second. Understanding that distinction is the starting point for any diagnosis that will actually help you.
What this article covers: the documented rules each platform applies, the three crawlers most people confuse, an ordered diagnostic you can run on your own site without purchasing anything, why the number you are using to judge this is probably measuring the wrong thing, and what no one in this space can honestly promise you.
What it does not cover: how to structure content for extraction once you know the answer is missing. That question is answered in depth at How to Structure a Knowledge Base for AI Search and Answer Engine Optimization Starts with Structure, which are the right places to go once this diagnostic tells you structure is the issue.
The short answer: these are two different eligibility tests
Ranking eligibility and citation eligibility are different tests, and a page can pass one while failing the other entirely.
Ranking eligibility is the question Google’s search index answers: can this page be returned for a query? It depends on crawlability, indexation, relevance signals, and authority. It is a retrieval question. The answer is binary: the page either enters the candidate pool for a query or it does not.
Citation eligibility is a different question that AI assistants answer independently: does this page contain a passage I can use to support the answer I am constructing? An assistant is not retrieving a list of ranked results and forwarding the top one. It is assembling an answer from multiple sub-questions, and for each sub-question it needs a passage that directly states the relevant fact, comparison, constraint, or definition. A page that ranks well but does not contain that specific passage will not be cited for that sub-question, regardless of its position.
| Ranking eligibility | Citation eligibility | |
|---|---|---|
| What it asks | Can this page be returned for a query? | Does this page contain a passage I can use in the answer I am building? |
| Who decides | Google’s search index | Each AI assistant, independently |
| What it depends on | Crawlability, indexation, relevance, authority | Crawler access, snippet and grounding directives, whether the answer is stated in a liftable form, entity clarity, outside corroboration |
| How it fails | The page never enters the candidate pool | The page is available but no passage answers the sub-question being resolved |
The gap between the two tests is where almost all the confusion lives. When a ranking page is not cited, the instinct is to treat it as an SEO problem and apply more of what worked for rankings: better metadata, more backlinks, updated content. Sometimes those things help. More often they do not, because the actual cause sits in a different layer of the system entirely.
The five most common causes, and the order in which to check them, are covered in the section below on the five reasons a ranking page is not cited. But the right starting point is the platform documentation itself, because it tells you what each system actually requires, and it contradicts several things currently being sold as solutions.
What Google actually says decides it, including the two things it says you do not need
Google’s position on AI Overviews and AI Mode eligibility is documented at Google Search Central, and it is simpler than most agency content suggests.
The requirement is this: “a page must be indexed and eligible to be shown in Google Search with a snippet.” That is the complete technical threshold. Google states plainly that “There are no additional technical requirements.” It also states that “There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.” And on structured data specifically: “There’s also no special schema.org structured data that you need to add.”
Those three statements eliminate a significant portion of what is currently being sold as AI optimization for Google’s surfaces. No proprietary schema, no special file, no separate submission process.
What the documentation does not say
“No additional requirements” is not the same as “nothing to do.” The requirement it names, snippet eligibility, is a real constraint that a meaningful number of sites fail without knowing it.
Snippet eligibility can be blocked in several ways. A nosnippet meta tag on a page tells Google not to show a text snippet for that result. The data-nosnippet attribute applied to a specific element tells Google not to extract text from that element. A max-snippet directive with a value of zero has the same effect as nosnippet. Any of these, applied intentionally or left over from a previous SEO decision, removes the page from AI Overview consideration without affecting its ranking position at all.
The practical implication is that a page can rank in position two and be completely invisible to AI Overviews because a snippet directive is blocking extraction. The ranking signal and the citation eligibility signal are controlled by different mechanisms, and they can point in opposite directions.
Check this before anything else. It is the cheapest check in the diagnostic, and it is the one most often skipped.
What OpenAI actually says decides it: three crawlers, and only one governs citation
ChatGPT’s citation behavior is governed by one specific crawler, and most sites that think they have opted out of ChatGPT have not done what they think they have done.
OpenAI’s bot documentation names three crawlers relevant to most sites:
- OAI-SearchBot surfaces websites in ChatGPT’s search features. This is the crawler that determines whether your content appears as a cited source in a ChatGPT search answer.
- GPTBot crawls content that may be used to train OpenAI’s foundation models. This is the training crawler.
- ChatGPT-User handles certain actions a user requests in ChatGPT or Custom GPTs. It is not used for automatic web crawling, and it does not determine whether content appears in Search.
These three crawlers are controlled independently. Blocking one does not block the others.
The error that removes sites from ChatGPT answers without anyone noticing
When publishers began blocking GPTBot to protect their content from being used in model training, many did so by adding a Disallow: / rule for GPTBot in their robots.txt file. That is a legitimate choice. The problem is that some sites, either through confusion or through broad bot-blocking rules, blocked OAI-SearchBot at the same time, or instead.
The documentation is direct: “Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers.” The same source states such sites “can still appear as navigational links,” but navigational links are not citations. They do not carry the same weight, and they do not constitute the kind of visibility that matters for the question this article is answering.
If you blocked GPTBot and your pages stopped appearing in ChatGPT answers, the robots.txt file is the first place to look. The two crawler names are different strings. A rule written for GPTBot does not affect OAI-SearchBot, and vice versa. But a wildcard rule, or a block applied to a broader user-agent pattern, can catch both.
One timing detail from the documentation that matters for diagnosis: “it can take ~24 hours from a site’s robots.txt update for our systems to adjust.” That is the only duration you need to remember. If you corrected a robots.txt block yesterday and are checking for results today, you are within the adjustment window. Wait the full period before concluding the fix did not work.
The practical implication is significant. A site can be fully indexed by Google, ranking well, with no snippet directives blocking extraction, and still be absent from ChatGPT answers because OAI-SearchBot is disallowed. That is a citation problem with a fast fix, and it has nothing to do with content quality, authority, or generative engine optimization.
The five reasons a ranking page is not cited, in the order to check them
These five causes account for the overwhelming majority of cases where a ranking page is absent from AI assistant answers. They are ordered cheapest and most decisive first: start at the top and rule each one out before moving to the next.
1. The search crawler is disallowed, or the page is not snippet-eligible
This is the access layer, and it is binary. If OAI-SearchBot is blocked in your robots.txt, ChatGPT cannot cite you regardless of anything else on this list. If Googlebot is blocked, or if a nosnippet directive is present, Google AI Overviews cannot cite you either.
Observable symptom: The page ranks but never appears as a cited source in any assistant answer on any query where it would be a natural fit.
The check: Fetch your robots.txt file directly at yourdomain.com/robots.txt and read it for OAI-SearchBot specifically, not just for Googlebot or a generic bot rule. Then inspect the page’s HTML for nosnippet, data-nosnippet, max-snippet:0, noarchive, or nocache directives. Both checks require no tools.
2. The page is indexed for a query nobody phrases that way
AI assistants do not receive the query your page ranks for. They receive the question the user typed, which may be phrased differently, and they decompose that question into sub-questions before retrieving sources. A page optimized for a precise commercial keyword may never surface during retrieval for a conversational prompt that covers the same topic, because the retrieval step uses different language.
Observable symptom: The page ranks for a specific keyword but is absent when you ask an assistant the underlying question in plain language.
The check: Ask the assistant the question the way a buyer would actually phrase it. Note which sources it cites. Then ask a follow-up that uses the language of your page’s target keyword. If the citation set changes, the retrieval gap is linguistic, not structural. The fix is ensuring your page also contains the plain-language version of the answer, not just the optimized version.
This is related to, but distinct from, answer engine optimization and conversational search optimization, both of which address how queries are phrased and matched.
3. The answer the reader wants is not stated on the page in a form that can be lifted
An assistant assembles its response from passages, not from pages. If the answer to the user’s question is implied by your page’s argument but never stated directly, the assistant cannot extract it. The information may be present in the sense that a careful reader would infer it. That is not sufficient. The passage needs to state the answer in a self-contained sentence or short paragraph that makes sense without the surrounding argument.
Observable symptom: The assistant’s answer is correct, and covers your topic, but cites a competitor whose page is less comprehensive than yours. The competitor’s page states the answer directly. Yours builds to it.
The check: Find the specific question the user is asking. Look for a sentence on your page that answers it directly, without requiring the reader to have read the preceding three paragraphs. If you cannot find that sentence, the answer is not liftable.
How to fix the structure once you have confirmed this is the cause is covered in detail at How to Structure a Knowledge Base for AI Search and Answer Engine Optimization Starts with Structure. This article stops at the diagnosis.
4. The entity is ambiguous, so the assistant cannot reliably attribute the page
AI assistants build answers around entities: organizations, products, people, locations. If the assistant cannot resolve which organization your page belongs to, or if the name used on your site differs from the name used in third-party sources, the assistant may deprioritize your page in favor of one where the entity is unambiguous.
Observable symptom: The assistant refers to your company or product by a different name, or describes it in terms that do not match your own positioning. Sometimes it omits you entirely in favor of a competitor with a more consistently named entity.
The check: Search for your company name across the sources the assistant is likely to use: industry directories, press coverage, LinkedIn, trade publications. Check whether the name, description, and category used in those sources matches what is on your site. Inconsistency across sources is an entity signal problem, not a content quality problem.
5. Nothing off your site corroborates the claim
AI assistants are more likely to cite a source when independent sources corroborate what it says. A claim that appears only on your own site, with no third-party reference to the same fact, carries less weight than the same claim supported by an external source. This is not about backlinks in the traditional SEO sense. It is about corroboration: does anything outside your domain confirm that this claim is accurate?
Observable symptom: Your page states a specific fact, figure, or position that is accurate but proprietary. The assistant either omits it or attributes it to a different source that has been cited by others.
The check: Identify the specific claim the assistant should be citing you for. Search for that claim across third-party sources. If nothing corroborates it, the assistant has no external signal to confirm it is accurate, and it will default to a source that something does corroborate.
One finding from an independent study is worth noting here, with its caveats stated plainly. Ahrefs analyzed 15,000 long-tail queries in early July 2025, submitting them to ChatGPT, Gemini, Copilot, and Perplexity. Across all four assistants, roughly 12% of the URLs cited ranked in Google’s top 10 for the original query. That average, however, is pulled upward significantly by Perplexity at 28.6%. The other assistants in the study sit nearer 8%. The data is from early July 2025, roughly sixteen months before this article’s publication, and the 15,000 queries were long-tail, so the figures should not be read as a universal benchmark. What the study does confirm is that ranking and citation are weakly correlated on every assistant measured, and most weakly correlated on the assistants that do not rely heavily on Google’s index. The five causes above explain the gap that those numbers describe.
How to tell which one is yours
Run these steps in order. Each one either rules a cause in or rules it out. Stop when you find the cause, fix it, then return to the list to confirm nothing else is also present.
Step 1: Check crawler access
Go to yourdomain.com/robots.txt in a browser. Read the file for the following user-agent strings: OAI-SearchBot, GPTBot, Googlebot. Look for Disallow: / or Disallow: /path-to-your-page under any of those user-agents. Also look for wildcard rules under User-agent: * that might be catching crawler traffic broadly.
If OAI-SearchBot is disallowed anywhere, that is your ChatGPT citation problem. The fix is removing or narrowing the disallow rule. Remember that ~24 hours is the documented adjustment window before the change propagates.
If Googlebot is disallowed for the page, that is your Google AI Overviews problem. Same fix, same principle.
For Bing’s surfaces, the Bing Webmaster Guidelines document additional directives that affect Copilot specifically. NOARCHIVE prevents content from being used in Copilot responses and grounding results. NOSNIPPET and DATA-NOSNIPPET may limit Copilot citation quality. NOCACHE restricts Copilot to the URL, title, and snippet, reducing citation depth. NOINDEX removes the URL from Bing search, Copilot experiences, and grounding results entirely.
| Directive | Platform | Effect |
|---|---|---|
noindex | Google, Bing | Removes the URL from search, and from Copilot and grounding results |
noarchive | Bing | Prevents the content being used in Copilot responses and grounding results |
nosnippet, data-nosnippet | Google, Bing | Google: removes the page from AI Overview consideration. Bing: prevents caption display and may limit Copilot citation quality |
max-snippet:0 | Same effect as nosnippet, and easy to miss | |
nocache | Bing | Limits Copilot to the URL, title and snippet, reducing citation depth |
Disallow on OAI-SearchBot | ChatGPT | Removes the site from ChatGPT search answers. Navigational links may remain |
Do not stop at the robots.txt. Also check whether your CDN, firewall, or bot-protection layer is blocking crawler traffic at the network level. A robots.txt that allows a crawler does not help if a firewall is returning a 403 before the crawler reaches the page.
Step 2: Check snippet eligibility
View the page source for the specific URL you are diagnosing. Search for the strings nosnippet, data-nosnippet, max-snippet, noarchive, and nocache. If any of these appear in the page’s meta robots tag or in the HTML of the page body, note which elements they apply to and whether the relevant answer content falls within a restricted block.
A max-snippet:0 in the meta robots tag is functionally equivalent to nosnippet and is easy to miss because it looks like a numerical value rather than a named directive.
Step 3: Confirm the page is indexed
Use the site: operator in Google to check whether the page is indexed at all: type site:yourdomain.com/your-page-path into Google and confirm the page appears. If it does not, indexation is the problem and the citation question is secondary. Request indexing via Google Search Console before continuing the diagnostic.
Step 4: Ask the assistant the question and read who it cites
Open ChatGPT, Gemini, or whichever assistant you are diagnosing. Ask the question your page should answer, phrased the way a real buyer would phrase it, not the way your page’s title tag phrases it. Read the full response. Note every source cited.
If your page is not among the sources, look at the pages that are. Ask yourself two questions. First: do those pages state the answer directly in a self-contained passage, while yours requires reading context to arrive at the same answer? Second: are those pages from organizations with more consistent third-party mentions than yours?
The answers to those two questions point directly to causes 3 and 5 from the previous section.
Step 5: Check entity consistency off-site
Search for your company name in quotation marks across Google, LinkedIn, and any industry directories or trade publications relevant to your sector. Note how the name, description, and category are presented in each source. If the name varies across sources, or if the description in external sources does not match your own positioning, you have an entity consistency problem.
The assistant is not reading your About page in isolation. It is cross-referencing what your site says against what other sources say about you. Inconsistency between those two signals reduces confidence in attribution, and reduced confidence in attribution reduces citation likelihood.
If you complete all five steps and the cause is still unclear, or if the diagnostic points to a structural issue with how the answer is presented on the page rather than an access or entity problem, the next step is a structured review of whether the answer is actually present and liftable. That review is the subject of Answer Engine Optimization Starts with Structure.
Why the number you are judging this on is probably wrong
Before acting on any citation measurement, it is worth understanding what that measurement can and cannot tell you. Three traps appear repeatedly, and each one can produce a confident-looking number that points in the wrong direction.
Trap 1: A single query is a sample of one, not a state
AI assistant outputs are non-deterministic. Ask the same question twice and you may receive two different answers with two different citation sets. Ask it ten times across a week and the set of cited sources may shift with each response, depending on when the assistant’s retrieval layer ran, what it found, and how the model weighted the results.
This means that checking whether your page appears in a ChatGPT answer on a given Tuesday tells you what happened on that Tuesday. It does not tell you your citation rate. It does not tell you whether you appeared yesterday or whether you will appear tomorrow. Treating a single observation as a state is one of the most common errors in how this topic is currently being measured and reported.
Trap 2: Citation share is an observational metric, not a ranking system
Bing Webmaster Tools introduced Citation Share reporting in its AI Performance section in June 2026. The Bing Webmaster blog post describes it clearly. Citation Share is “the percentage of citations attributed to your site out of all citations shown across all sites.” It is “designed as an observational metric – not a ranking system or a competitive scoreboard.” It “does not expose competitor domains, represent traffic share, or assign quality scores to content.” And the tools that report it “do not turn AI visibility into a single ranking or score.”
That is Microsoft’s own characterization of its own measurement. An observational metric tells you what happened. It does not tell you why, and it does not tell you what to change. Using Citation Share as a performance target, rather than as a diagnostic signal, is using the instrument outside its documented purpose.
Trap 3: Impression counts can include non-human querying
Any system that measures how often your site appears in AI assistant answers needs to send queries to those assistants to observe the outputs. When queries are sent at scale, by automated systems rather than by individual people, the impression count includes those automated queries alongside real user queries. The two are not separated in the reported figures.
This does not invalidate measurement as a practice. It does mean that volume figures and apparent position data should be read with awareness that the denominator may include automated traffic that does not represent actual buyer intent. A high impression count in a tracking tool is not the same as high visibility among the buyers you are trying to reach.
None of this means measurement is useless. It means the measurement needs to be understood for what it is: a signal, not a verdict. The right use of these tools is to identify patterns over time, not to draw conclusions from a single data point.
What none of this promises
Working through this diagnostic and fixing every issue it surfaces does not guarantee a citation. That is not a hedge. It is what the platform documentation states.
Google is direct on this point: “Indexing and serving isn’t guaranteed.” A page can meet every documented requirement and still not appear in a given AI Overview, because the system decides in real time whether to include a source, which sources to include, and how to weight them. There is no submission process, no approval queue, and no confirmation that a compliant page will be used.
The same logic applies to ChatGPT, Gemini, and Copilot. Allowing OAI-SearchBot removes a barrier. It does not create an entitlement. Fixing an entity consistency problem makes attribution more reliable. It does not make citation certain.
What this diagnostic does promise is something more useful than a citation guarantee. It tells you whether the barriers that are within your control have been removed. If crawler access is clean, snippet eligibility is confirmed, the page is indexed, the answer is present and liftable, the entity is consistent, and corroborating sources exist, then the remaining variable is the assistant’s real-time judgment. That is not something any advisor can control, and any advisor who claims otherwise is selling something the documentation does not support.
What you should be able to do after running this diagnostic is tell the difference between a fixable technical problem and a platform decision. Those are different situations that call for different responses. Conflating them is expensive, because the fixes for technical problems are cheap and the fixes for platform decisions do not exist.
If the diagnostic surfaces issues you cannot resolve in-house, or if the five-step process points to causes that require a more structured review of your site’s technical configuration and content architecture, a second opinion from someone who works from documentation rather than from sales claims is worth the conversation. The AI Visibility Audit is where that conversation starts.
The reader who finishes this article should be able to do one thing they could not do before: run a diagnostic on a specific page, identify the most likely cause of a citation gap, and tell a real technical problem from a pitch. That is the capability this article is designed to transfer. Whether you run it yourself or bring in help, the sequence is the same and the documentation is public. Start with the robots.txt.
Run the diagnostic on your own site, or have us run it.
We check crawler access, snippet and grounding directives, whether your answer is present in a liftable form, and what outside sources say your company is. You get back what is blocking citation, what is fixable this week, and what is a platform decision nobody can fix.
See what an AI Visibility Audit covers Already cited, but described wrongly? That is a different job. Answer Engine Optimization.Frequently Asked Questions
What should I do if an AI assistant cites a competitor’s page that is thinner than mine?
The assistant is not comparing page depth. It is selecting the passage that most directly answers the specific sub-question it is trying to resolve. If a thinner competitor page states the answer in one self-contained sentence and your page builds to the same answer over three paragraphs, the thinner page wins that sub-question. The fix is not to add more content to your page. It is to confirm that your page also contains a direct, self-contained statement of the answer, so the assistant has something liftable without reading surrounding context. The structure question, once you have confirmed the answer is absent or buried, is covered at How to Structure a Knowledge Base for AI Search.
If I remove the robots.txt block on OAI-SearchBot today, how long before my pages can appear in ChatGPT answers?
OpenAI’s documentation states that it can take ~24 hours from a robots.txt update for their systems to adjust. That is the only duration the documentation names. It is a propagation window, not a citation promise. Once the window closes, the crawler can access the page, but whether the assistant then selects it as a cited source depends on the other four causes in the diagnostic. Crawler access is necessary. It is not sufficient.
Can a page be cited by an AI assistant without being indexed by Google?
Yes, for assistants that do not rely on Google’s index. ChatGPT Search uses OAI-SearchBot to crawl the web independently. Perplexity operates its own crawler. Neither requires Google indexation as a prerequisite. Google AI Overviews and AI Mode are the specific surfaces where Google’s own indexation requirement applies, because those surfaces are built on Google Search eligibility. If your concern is Google AI Overviews specifically, indexation is the threshold. If your concern is ChatGPT or Perplexity, the relevant question is whether those platforms’ crawlers can access and retrieve the page.
What is the difference between a navigational link and a citation in a ChatGPT answer?
OpenAI’s documentation draws this distinction directly. When a site opts out of OAI-SearchBot, it will not appear as a cited source in ChatGPT search answers, but it “can still appear as navigational links.” A navigational link is a reference that helps a user find a resource. A citation is a source the assistant used to construct its answer. The two carry different weight. A navigational link does not indicate that the assistant read and selected your content. It does not transfer authority in the way a citation does, and it does not constitute the kind of AI visibility that most businesses are trying to measure.
What happens when two different AI assistants give contradictory information about who my company is?
Each assistant builds its understanding of an entity from the sources it retrieves and the signals it weights. If your company name, description, or category is presented inconsistently across third-party sources, different assistants may resolve the ambiguity differently, and neither resolution may match your own positioning. The practical consequence is that one assistant may describe your company accurately while another uses a different name, a different category, or a description that reflects an older version of the business. The fix is not assistant-specific. It is off-site entity consistency: the name, description, and category that appear in industry directories, press coverage, LinkedIn, and trade publications should match what your site says. Inconsistency across those sources is what creates the divergence.
Does updating the page change what the assistant says immediately?
No. An assistant’s answer reflects what it retrieved the last time it crawled and indexed the relevant sources, not what the page says right now. The lag between a page update and a change in an assistant’s answer depends on how frequently the assistant’s crawler revisits the page, how quickly the updated content enters the retrieval index, and whether the updated passage is then selected for a given answer. Google documents no submission process for its AI surfaces, and none of the major assistants offers one publicly. The practical implication is that citation behavior should be assessed against the page as it existed during the crawler’s last visit, not the page as it exists today.
If only one assistant omits me while others cite me, what should I check first?
Start with crawler access for that specific assistant. Each platform operates its own crawler under a different user-agent string. A robots.txt rule that blocks one assistant’s crawler does not affect the others, and a wildcard rule may catch one while leaving others unaffected. Fetch your robots.txt and read it for the specific user-agent of the assistant that is omitting you. If crawler access is clean, the next question is whether that assistant’s retrieval system has reached your page at all. Assistants weight and retrieve sources differently, so one may simply not have retrieved your page yet while another has. Check whether the assistant that omits you also omits other pages on your site, or whether the gap is specific to the page you are diagnosing.
Is there a way to check whether OAI-SearchBot has actually visited my site?
Yes. Your server access logs record every crawler visit by user-agent string. Search the logs for OAI-SearchBot. If you find recent visits, the crawler is accessing the site. If you find no visits, either the crawler has not yet reached the site, or something upstream is blocking it before the request reaches the server, such as a firewall rule, a CDN configuration, or a bot-protection layer that does not honor the robots.txt allowance. A robots.txt that permits OAI-SearchBot does not help if a network-level rule is returning a 403 before the crawler reaches the page. If you cannot access server logs directly, your hosting provider or CDN dashboard typically surfaces bot traffic by user-agent. If a structured review of your technical configuration is what the diagnostic points to, the AI Visibility Audit is where that work starts.
Can structured data schema improve citation likelihood if everything else is already in order?
Google’s documentation is direct on this: “There’s also no special schema.org structured data that you need to add” to appear in AI Overviews or AI Mode. That is a statement about requirements, not about correlation. Schema may improve how Google parses and presents page content in other contexts, and it remains useful for other reasons. But if crawler access is clean, snippet eligibility is confirmed, the page is indexed, the answer is liftable, the entity is consistent, and corroborating sources exist, adding schema is unlikely to change citation behavior on Google’s surfaces. Treat schema as hygiene, not as a citation lever.
How do I know if the diagnostic has identified the real cause or just the most visible one?
Run all five steps before acting on any single finding. It is possible for a page to have more than one cause operating at the same time. A site that blocked OAI-SearchBot and also has a nosnippet directive and also has entity inconsistency off-site has three separate problems, and fixing only the most visible one will not produce the result the fix appeared to promise. The diagnostic is ordered cheapest and most decisive first because that order minimizes wasted effort, not because each cause is mutually exclusive. Once you have fixed a cause, return to the full list and confirm nothing else is also present.