Chain of Title and Title Examination

Chapter 33 ended with a statute and a concept. The statute told us which of two claimants a State prefers; the concept of notice told us who is disqualified from that preference. Neither told us what a purchaser is required to look at. That is the subject of this chapter, and it is the point at which recording law stops being a set of rules and becomes a professional method.

The organizing proposition is this. Constructive notice is not a function of what the records contain; it is a function of what a legally sufficient search would reveal. The chain of title is the name we give to the set of instruments that such a search discloses. Everything that follows — wild deeds, early and late recordings, common-grantor restrictions, indexing errors, after-acquired title, curative acts, marketable record title — is an elaboration of the distance between what is in the books and what is in the chain.

1. The chain of title is defined by the search, not by the archive. An instrument imparts constructive notice only if the search method the law requires of a purchaser would disclose it.
2. A chain of title is a reconstruction, not a document. The examiner builds it link by link from the indexes and then reads each instrument to test whether the link holds.
3. The grantor-grantee indexes are searched in two directions. A backward search from the present owner establishes the sequence of grantors; a forward search from each owner tests whether that owner conveyed to anyone else while holding title.
4. Each owner must be searched from the date title vested until the date the owner's conveyance was recorded. The period of exposure, not the period of ownership, defines the run sheet.
5. A tract index, where it exists, searches the land rather than the parties. It cures many chain problems but is available in a minority of jurisdictions and does not displace instrument reading.
6. An instrument recorded before the grantor acquired title is often outside the chain. The early-recorded deed problem is jurisdictionally split and is closely bound to estoppel by deed.
7. An instrument recorded after the grantor has conveyed away is often outside the chain. The late-recorded deed problem, illustrated by the classic Massachusetts and Minnesota decisions, turns on how far forward a searcher must run each grantor.
8. A wild deed imparts no constructive notice. A deed whose grantor never appears as a grantee of record is unconnected to the chain and cannot be found by a proper search.
9. Estoppel by deed and after-acquired title operate between the parties with certainty and against third parties only imperfectly. The doctrine passes the later-acquired estate to the earlier grantee, but the recording system may still prefer a subsequent bona fide purchaser.
10. Restrictions imposed by a common grantor on other parcels present the hardest chain question in American law. Jurisdictions divide on whether deeds out of a common grantor to neighboring lots are within a purchaser's chain.
11. Inquiry notice reaches beyond the record and is not defeated by a clean search. Possession, visible use, recitals, and references in recorded instruments impose duties of investigation regardless of what the indexes disclose.
12. A recital in a recorded instrument charges the purchaser with what reasonable pursuit of it would reveal. The examiner reads instruments; the examiner does not merely list them.
13. Indexing errors produce a genuine jurisdictional split. Some States hold that proper submission is enough; others hold that an instrument the indexes cannot reveal imparts no notice.
14. Name variance, marital-status changes, entity conversions, and misspellings break chains that are substantively sound. Idem sonans, corrective deeds, affidavits, and scrivener's-error statutes are the professional response.
15. Void links cannot be cured by recording or by good faith; voidable links can. Forgery, non-delivery, and want of capacity destroy the link; fraud in the inducement and similar defects yield to a bona fide purchaser.
16. The shelter rule carries protection forward through the chain. A transferee of a protected purchaser takes that protection even with notice, subject to the rule barring a wrongdoer from reacquiring.
17. Three-party and four-party problems are solved pairwise and in sequence. Priority is determined between each adjacent pair under the governing statute and then composed, with the shelter rule applied last.
18. Judgment liens, tax liens, mechanics' liens, and lis pendens are searched separately from the deed indexes. Their attachment rules and duration are statutory and are examined as part of title, not as an afterthought.
19. Probate, guardianship, receivership, and fiduciary links must be tested against the authority of the transferor. A deed from a personal representative is only as good as the order, license, or power that authorized it.
20. Curative acts and marketable record title legislation are the system's self-repair mechanisms. Curative statutes validate defective instruments after a period; marketable-title acts extinguish stale interests measured from a root of title unless preserved.

1. Define the chain of title functionally, as the product of a legally sufficient search, and explain why that definition governs constructive notice.
2. Construct a chain of title from grantor-grantee indexes by backward and forward searching, and record it on a run sheet.
3. State, for each owner in a chain, the precise period during which that owner must be searched as a grantor.
4. Explain the operation and limits of a tract index and compare it with name indexing.
5. Distinguish instruments inside the chain from instruments recorded too early, too late, or wholly outside it.
6. Analyze the wild-deed problem and explain why physical presence in the records is immaterial.
7. Apply estoppel by deed and after-acquired title, and explain the competing treatment of a subsequent purchaser.
8. Analyze the common-grantor problem and identify the three principal jurisdictional approaches.
9. Identify the facts that trigger inquiry notice and describe the scope of the inquiry the law requires.
10. Analyze indexing errors, name variance, and the curative instruments that respond to them.
11. Distinguish void from voidable links and trace the consequences through the chain.
12. Apply the shelter rule in multi-party sequences.
13. Resolve three-party and four-party priority problems by pairwise sequencing.
14. Search and evaluate judgment liens, tax liens, mechanics' liens, and lis pendens as objects of title examination.
15. Test probate, fiduciary, and entity links for authority.
16. Explain the operation of curative acts and marketable record title legislation, and their limits.
17. Execute a professional title-examination methodology from the search order to the written opinion, and identify the exceptions and requirements it must state.

• Board of Education v. Hughes, 136 N.W. 1095 (Minn. 1912) (deed recorded before the grantor's title appeared of record)
• Ayer v. Philadelphia & Boston Face Brick Co., 34 N.E. 177 (Mass. 1893) (late-recorded deed outside the chain)
• Morse v. Curtis, 2 N.E. 929 (Mass. 1885) (limits of the forward search)
• Woods v. Garnett, 16 So. 390 (Miss. 1894) (contrary view on late recording)
• Sabo v. Horvath, 559 P.2d 1038 (Alaska 1976) (after-acquired title and the subsequent purchaser)
• Guillette v. Daly Dry Wall, Inc., 325 N.E.2d 572 (Mass. 1975) (deeds out of a common grantor)
• Sanborn v. McLean, 206 N.W. 496 (Mich. 1925) (reciprocal negative easement and inquiry notice)
• Witter v. Taggart, 577 N.E.2d 338 (N.Y. 1991) (deeds out of a common grantor; contrary approach)
• Luthi v. Evans, 576 P.2d 1064 (Kan. 1978) (blanket description and constructive notice)
• Messersmith v. Smith, 60 N.W.2d 276 (N.D. 1953) (defective acknowledgment; record imparts no notice)
• Mortensen v. Lingo, 99 F. Supp. 585 (D. Alaska 1951) (instrument outside the chain)
• Waldorff Insurance & Bonding, Inc. v. Eglin National Bank, 453 So. 2d 1383 (Fla. Dist. Ct. App. 1984) (possession and inquiry notice)
• Miller v. Green, 58 N.W.2d 704 (Wis. 1953) (possession as notice)
• Harper v. Paradise, 210 S.E.2d 710 (Ga. 1974) (recitals, missing instruments, and duty of inquiry)
• Model Marketable Title Act (1960) and enacting statutes (e.g., Fla. Stat. §§ 712.01–712.12; Mich. Comp. Laws §§ 565.101–565.109; Ohio Rev. Code §§ 5301.47–5301.56)
• Uniform Simplification of Land Transfers Act arts. 3–4 (Unif. L. Comm'n 1976)
• Uniform Real Property Electronic Recording Act (Unif. L. Comm'n 2004, rev. 2005)
• Representative curative statutes (e.g., Mass. Gen. Laws ch. 184, § 24; Cal. Civ. Code § 1207)
• Representative lis pendens statutes (e.g., Cal. Code Civ. Proc. § 405.20; N.Y. C.P.L.R. 6501)
• Restatement (Third) of Property: Servitudes §§ 2.14, 7.14 (Am. L. Inst. 2000)

Statutory citations are representative examples of drafting patterns. Search periods, curative provisions, lien durations, and marketable-title roots are creatures of local law and must be verified against the governing statutes and decisions of the situs before any examination is relied upon.

• Patton and Palomar on Land Titles §§ 12–70 (3d ed.) (chain of title, search practice, and examination standards)
• 4 American Law of Property §§ 17.17–17.29 (A. James Casner ed. 1952)
• 11 Richard R. Powell, Powell on Real Property chs. 82–84, 92
• 4 Herbert Thorndike Tiffany, The Law of Real Property §§ 1259–1318 (3d ed.)
• 14 Thompson on Real Property (Thomas ed.) chs. 92–93
• Lewis M. Simes & Clarence B. Taylor, The Improvement of Conveyancing by Legislation (1960)
• Walter E. Barnett, Marketable Title Acts — Panacea or Pandemonium?, 53 Cornell L. Rev. 45 (1967)
• Paul E. Basye, Clearing Land Titles (2d ed. 1970)
• John L. McCormack, Torrens and Recording: Land Title Assurance in the Computer Age, 18 Wm. Mitchell L. Rev. 61 (1992)
• American Land Title Association, Title Insurance Commitment and Policy Forms (as revised) (for the examination-to-assurance interface treated in Chapter 35)

A chain of title is the ordered succession of recorded instruments by which an estate in a particular parcel is traced from a recognized starting point to the present record owner. It is not a document held anywhere. It is a reconstruction, performed by an examiner, from indexes that are organized by the names of parties rather than by land.

The reconstruction has two components, and confusing them is the most common error of beginners. The first is the skeleton: a list of conveyances, each linking a grantee who becomes the next grantor. The second is the substance: the reading of each instrument in that list to determine what estate it purported to convey, on what conditions, with what reservations, exceptions, recitals, and references. A chain that is complete on the index cards may be worthless once the instruments are read, and a chain that appears broken may be sound once the instruments explain the break.

The legal significance of the chain is that constructive notice is measured by it. Chapter 33 established that a recorded instrument imparts notice only if the record was one the law requires the purchaser to consult. This chapter supplies the content of that requirement. A purchaser is charged with everything a properly conducted search would reveal, and with nothing the records happen to contain outside that search, subject always to the independent operation of actual and inquiry notice.

It follows that the phrase “in the chain of title” is a conclusion, not a description. To say that an instrument is in the chain is to say that the search method the law prescribes in that jurisdiction would have produced it. Because search methods differ — grantor-grantee against tract, longer against shorter periods, broader against narrower forward runs — an instrument may be inside the chain in one State and outside it in the State across the river, on identical facts.

Every search must begin somewhere, and the choice of a starting point is a legal question rather than a matter of convenience. The theoretical ideal is a search to the sovereign: to the original patent, grant, or crown conveyance from which all private title in the parcel derives. In a few jurisdictions with short recorded histories that ideal remains attainable, and in some institutional contexts it is still required.

In practice the search period is bounded in one of three ways. First, by a marketable record title act, which supplies a statutory root of title — typically the most recent conveyance recorded at least thirty or forty years before the search — and extinguishes most interests arising before it unless preserved by a recorded notice. Second, by local title standards or bar association custom, which may fix a sixty-year, fifty-year, or forty-year period as a professional norm. Third, by the terms of an examination engagement or a title insurer's underwriting requirements.

The examiner must know which of these governs, because each has a different legal consequence. A marketable-title act truncates the search with statutory force: interests before the root are extinguished, not merely unexamined. A custom or standard truncates the search as a matter of professional practice only: an interest before the customary period remains legally effective, and the examiner who ignores it relies on the improbability of its assertion, on curative legislation, or on insurance.

The distinction is the difference between a title that is good and a title that is merely customary. Chapter 35 develops the institutional response to that gap. Here it is enough to say that the examiner must state the period searched, and must state it as an express limitation on the opinion rendered.

American land records are, with limited exceptions, indexed by name. The grantor index lists instruments alphabetically by the name of the party conveying; the grantee index lists them by the name of the party receiving. Each entry gives the parties, the instrument type, the date of recording, a book and page or instrument number, and often an abbreviated description. The index entry is a finding aid; the recorded instrument is the legal document, and the two must never be confused.

A minority of jurisdictions maintain a tract index, which organizes entries by parcel: by section, township, and range in public-land-survey States, or by lot and block in platted subdivisions. Where a tract index exists and is legally authoritative, the search problem is transformed, because instruments affecting the land appear in one place regardless of the names of the parties. Wild deeds become discoverable; late-recorded deeds appear on the parcel's page; common-grantor restrictions surface without a separate theory.

Two cautions attach. First, in many States the tract index is maintained as a convenience and is not the statutory index; constructive notice is measured by the name indexes, and the tract index is an aid without legal force. Second, even an authoritative tract index resolves discovery only. It does not tell the examiner whether the instrument was validly executed, whether it was delivered, whether the grantor had authority, or what its language means.

Digital search has changed the practical texture of examination without changing this framework. An electronic index may permit searching by name, instrument number, legal description, and sometimes full text, and may return results a manual search would miss. Constructive notice, however, remains measured by the search the law prescribes, not by the search the software makes possible. A jurisdiction that measures notice by proper name indexing does not enlarge it merely because a full-text query would have found a misindexed deed.

The examiner begins with the present record owner and works backward through the grantee index. The question at each step is: from whom did this owner take, and when was that conveyance recorded? The answer identifies the prior owner, who becomes the subject of the next query, and the process repeats until the search reaches the root of title or the sovereign.

The product is a run sheet: a table of links, each showing grantor, grantee, instrument type, execution date, recording date, and book and page. The run sheet is the skeleton of the chain, and it should be built completely before any instrument is read in detail, because an early misreading can send the examiner down a chain that belongs to a different parcel or a different family of the same surname.

Three problems recur in the backward search. The first is the missing link: the grantee index discloses no conveyance into a person who nonetheless appears as a grantor. The gap may be explained by devise, intestate succession, a decree, a name change, a marital-status change, a corporate merger, or simply by a failure to record. Each explanation requires its own proof, and probate, court, and corporate records are searched accordingly.

The second is the parcel-identity problem: the described land changes shape across the chain by subdivision, assemblage, partial release, taking, or boundary agreement. The examiner must trace the parcel as well as the title, and must be certain that the earlier deed's description encompasses the land now examined.

The third is name variance. Titles are broken by middle initials, married names, transliterations, misspellings, and entity conversions far more often than by fraud. The doctrine of idem sonans and local search conventions mitigate the problem, but the professional response is to search plausible variants and to document the reasoning.

The backward search establishes who owned the land. It cannot establish that those owners did not also convey it, or encumber it, to someone else. That is the office of the forward search, and it is where the intellectual content of the recording system lies.

For each owner in the chain, the examiner returns to the grantor index and runs that owner forward through a defined period, examining every instrument in which the owner appears as a grantor and determining whether it affects the parcel. Mortgages, easements, leases, restrictive covenants, options, contracts of sale, releases, and second deeds are all disclosed here.

The period to be run is the heart of the matter. The conventional and safest rule runs each owner from the date title vested in that owner — not from the date the deed to that owner was recorded — until the date the deed out of that owner was recorded. This period captures the two intervals during which a competing instrument may be created: the gap between delivery and recording of the deed in, and the gap between delivery and recording of the deed out.

Some jurisdictions permit a narrower run, from recording of the deed in to recording of the deed out, and a few limit the forward run further under decisions such as Morse v. Curtis. Others require a broader run, extending past the recording of the deed out to capture late-recorded conveyances. The examiner must know the local rule, because the difference is not stylistic: it determines whether whole categories of instruments are inside or outside the chain.

A chain assembled from index entries is a hypothesis. It becomes an examination only when each instrument is read. The examiner reads for the granting clause and the estate conveyed, for the description and its consistency with the parcel, for reservations and exceptions, for conditions and executory limitations, for covenants of title, for the identity and capacity of the parties, for acknowledgment and attestation, and above all for recitals and references.

Recitals are the connective tissue of a chain. A deed that recites that the grantor takes as devisee under a named will, or as surviving joint tenant, or subject to a described mortgage, or that refers to an unrecorded agreement, does two things at once. It supplies the examiner with a lead, and it charges the purchaser with the duty to pursue it. Harper v. Paradise is the classic instance: a deed referring to an earlier, missing instrument put subsequent takers on notice of the interest that instrument created.

References to plats, surveys, and prior instruments have the same effect. A conveyance of “Lot 14 as shown on the plat recorded in Plat Book 3, Page 22” incorporates the plat, and the purchaser is charged with the easements, setbacks, dedications, and restrictions the plat discloses. Likewise, a deed “subject to restrictions of record” does not identify the restrictions but confirms their existence and directs the search.

The lesson is procedural as much as doctrinal. A title examination is a reading exercise embedded in a search. The examiner who lists instruments without reading them will produce a chain that is complete and an opinion that is wrong.

The category of instruments that are recorded and yet impart no constructive notice is the recording system's most important qualification. It exists because notice is measured by the required search, and a search organized by names cannot reveal instruments the names do not connect.

Four principal categories recur. The wild deed, whose grantor never appears of record as a grantee, is unconnected at its root. The early-recorded deed, made and recorded before the grantor acquired title, appears in the grantor index under a period the searcher has no reason to examine. The late-recorded deed, recorded after the grantor has apparently conveyed the land away, appears after the point at which the forward run ordinarily stops. The deed out of a common grantor to a neighboring parcel, containing restrictions purporting to bind the parcel retained, appears in the grantor index but under a conveyance of other land.

To these should be added instruments that are unrecordable for want of acknowledgment yet physically present in the books, treated in Chapter 33 under Messersmith, and instruments whose descriptions are too indefinite to identify the land, as in Luthi v. Evans, where a blanket assignment of “all interests in the county” was held not to impart constructive notice as to a specific parcel.

In each case the analysis is identical and should be stated in the same form: would the search the law requires have disclosed this instrument? If not, the instrument does not charge the purchaser with record notice, and the purchaser's protection then depends only on whether actual or inquiry notice exists independently.

A wild deed is an instrument whose grantor is not, on the record, a grantee. Suppose that O owns Blackacre of record. X, a stranger to the record, purports to convey Blackacre to A, and A records. Later O conveys Blackacre to B, who searches the records. B runs the grantee index back from O, finds O's title regular, runs each owner forward, and finds nothing. The X-to-A deed is invisible, because a searcher has no reason to look under X's name at all.

The rule is that such a deed imparts no constructive notice. It is recorded, but it is not in B's chain, and B is charged only with what the chain discloses. A's remedy, if any, lies against X on the covenants of title, not against B.

The doctrine hardens when the wild deed is later connected. If O afterward conveys to X, and X's earlier deed to A is already of record, the chain is completed retroactively in a substantive sense but not in a search sense: the deed from X to A remains indexed under a period no searcher would run. This is precisely the early-recording problem, and the classic case is Board of Education v. Hughes, holding that a deed recorded before its grantor's title appeared of record does not give constructive notice to a subsequent purchaser.

The professional response is simple and should be stated as a rule of practice: when a grantor's title is acquired after the grantee's deed is recorded, the deed should be re-recorded once the grantor's acquisition appears of record. Re-recording places the instrument within the period the searcher will run and converts an invisible instrument into a visible one.

The early-recorded deed and the late-recorded deed are mirror problems, and both turn on the boundaries of the forward run. The early deed is recorded before the grantor's own acquisition appears; the late deed is recorded after the grantor has conveyed the land away.

Late recording produces the classic Massachusetts rule. In Ayer v. Philadelphia & Boston Face Brick Co. and in Morse v. Curtis, the court declined to charge a purchaser with notice of an instrument recorded after the grantor's deed out had been recorded, on the ground that a searcher must be permitted to stop running each grantor at some determinate point. The competing view, associated with Woods v. Garnett and adopted in a number of States, requires the searcher to run each grantor forward indefinitely, on the ground that the burden of the further search is smaller than the injustice of defeating a recorded interest.

Neither position is unprincipled. The Massachusetts rule optimizes for the determinacy of the search and accepts that some recorded interests will be defeated. The contrary rule optimizes for the protection of recorded interests and accepts that searches become longer and less certain. The examiner's obligation is to know which rule governs and to search accordingly; the drafter's obligation is to record promptly and thereby to avoid the question.

The same reasoning explains why prompt recording is the single most effective act of title hygiene available to a grantee. Every chain-of-title pathology in this Part is a consequence of a delay between conveyance and recordation, or of a recordation made at a moment when the records could not accommodate it.

A developer owns twenty lots. Selling lots one through nineteen, the developer inserts in each deed a covenant restricting all twenty lots, including the retained lot twenty, to residential use. The developer then conveys lot twenty without restrictions to a purchaser who searches the records. Is the purchaser charged with the restrictions recited in the deeds of the other nineteen lots?

The question is genuinely hard because the restrictions are recorded, and they are recorded under the developer's name as grantor during the period the searcher must run. What they are not recorded under is the parcel being purchased. The searcher who reads only instruments describing lot twenty will not find them; the searcher who reads every instrument in which the developer appears as grantor will.

Three approaches divide the States. The Massachusetts approach, exemplified by Guillette v. Daly Dry Wall, charges the purchaser with the restrictions, reasoning that deeds out of the common grantor during the period of ownership are within the chain and must be examined. The New York approach, exemplified by Witter v. Taggart, refuses to charge the purchaser, reasoning that a restriction affecting a parcel must appear in that parcel's own direct chain. A third group reaches the purchaser through inquiry notice rather than record notice, as in Sanborn v. McLean, where the visibly uniform residential character of a neighborhood was held to put a purchaser on inquiry that would have revealed a common scheme.

For the examiner, the practical rule is disjunctive. In common-grantor jurisdictions the deeds out of a common grantor must be examined; in all jurisdictions the physical character of the neighborhood, the presence of a recorded plat, and any declaration of restrictions must be checked. For the drafter, the answer is to record a declaration of restrictions describing all burdened parcels before the first conveyance, which places the restriction in every parcel's direct chain and moots the dispute.

A clean record search never establishes that a purchaser took without notice. It establishes only that the purchaser took without record notice. Actual knowledge and inquiry notice operate independently, and inquiry notice is the principal reason that examination cannot be conducted from a desk alone.

Inquiry notice arises when the purchaser knows facts that would cause a reasonable person to investigate. The law then charges the purchaser with whatever a reasonable investigation would have revealed, whether or not it was undertaken. The classic triggers are four: possession inconsistent with the record title; visible physical conditions such as paths, utility lines, drainage, and party walls; recitals and references in recorded instruments; and unexplained irregularities in the chain itself, such as a nominal consideration, a quitclaim in an otherwise warranty chain, or a rapid resale at a disproportionate price.

Possession is the strongest of these. In Miller v. Green and in Waldorff, occupancy by someone other than the record owner was held to charge a purchaser with the occupant's rights. The purchaser is expected to inspect, to ask the occupant by what right the occupant holds, and to accept the consequences of the answer. A purchaser who does not inspect is in the same position as one who inspected and ignored what was seen.

The scope of the inquiry matters as much as its trigger. The law does not require an unlimited investigation; it requires the investigation a reasonable person would pursue and charges the purchaser with what that investigation would produce. An inquiry that is begun and abandoned without reason charges the purchaser fully; an inquiry pursued diligently and frustrated by the occupant's refusal to answer may, in some jurisdictions, discharge the duty.

An instrument correctly submitted and correctly recorded may nonetheless be misindexed by the recorder, or indexed under a name that a diligent searcher would not consult. The question is who bears the loss: the grantee who did everything the statute required, or the later purchaser who searched as the law prescribes and found nothing.

The States divide. One line holds that the grantee's obligation ends with proper submission, so that a properly filed but misindexed instrument imparts constructive notice; the loss falls on the searcher, whose recourse is against the recorder. The other line holds that constructive notice depends on discoverability, so that an instrument the indexes do not reveal imparts no notice; the loss falls on the grantee, whose recourse is likewise against the recorder. A number of States have addressed the allocation by statute, and several make indexing an express element of recording.

Name variance is a related but distinct pathology, and in volume it is far more significant. Titles break on middle initials, on maiden and married names, on anglicized and transliterated spellings, on trade names, on corporate successions and conversions, and on the many ways in which trustees and fiduciaries are described. The doctrine of idem sonans treats names that sound alike as equivalent where no one is misled, but it is a doctrine of last resort rather than a search strategy. The professional response is to search variants systematically and to document them.

Where a defect is discovered, the corrective apparatus is well developed: the corrective or confirmatory deed, executed by the original parties to restate the conveyance accurately; the re-recording of an instrument once a prior defect is cured; the affidavit of identity, heirship, or scrivener's error, admissible in many States to explain a break; the quitclaim from a person whose apparent interest clouds the title; the disclaimer; and, where the parties will not cooperate, the action to quiet title or to reform the instrument. Each is a different tool with different formal requirements, and the examiner selects among them by the nature of the defect.

A chain is only as strong as each link, and links fail in two fundamentally different ways. A void link conveys nothing and cannot be validated by recording, by good faith, or by the passage of a purchaser through the chain. Forgery is the paradigm: a forged deed transfers no estate, and every conveyance derived from it is equally ineffective, however innocent the later takers. Non-delivery, a deed from a wholly incompetent grantor, a deed from a non-existent grantee, and a conveyance by one with no interest at all belong to the same category.

A voidable link, by contrast, conveys the estate subject to a power of avoidance in the defrauded or imposed-upon party. Fraud in the inducement, undue influence, breach of fiduciary duty, and defects short of incapacity produce voidable conveyances. The rule of long standing is that the power of avoidance is cut off by a bona fide purchaser for value without notice: the defrauded grantor may rescind against the fraudulent grantee, but not against the innocent purchaser who takes from that grantee before rescission.

The distinction is doctrinally decisive and practically difficult, because the two situations look identical in the record. Neither a forged deed nor a fraudulently obtained deed announces itself; both appear as regular conveyances. The examiner's tools are indirect: irregular acknowledgments, notaries whose commissions had expired, conveyances by persons then known to be deceased or incapacitated, unexplained conveyances out of the ordinary pattern, and transfers for nominal consideration. Where these appear, the examiner raises a requirement rather than an opinion.

Chapter 33 stated these rules in outline. Their significance here is that the examiner encounters them as links, not as abstractions, and must decide whether the chain can be certified at all. A void link is not an exception to be listed; it is a break that must be repaired or the title declined.

A grantor who purports to convey an estate not then owned, and who warrants the title conveyed, is estopped as against the grantee to deny the conveyance when the estate is afterward acquired. The estate is treated as passing automatically to the earlier grantee at the moment of acquisition. The doctrine is called estoppel by deed, and its product is called after-acquired title.

Between the parties the rule is settled and generous. It applies most clearly where the deed contains covenants of general warranty or an express representation of the estate conveyed, and it is ordinarily denied to a grantee under a bare quitclaim, which purports to convey only whatever the grantor has. Some jurisdictions extend it where the deed recites a specific estate even without warranties.

The recording problem arises when a third party intervenes. Suppose O has no title; O conveys to A by warranty deed, and A records. O then acquires title from the true owner and conveys to B, who searches, finds O's acquisition and O's conveyance, and finds nothing to suggest the earlier deed to A — because the deed to A was recorded before O's title appeared of record and is therefore outside B's chain. Estoppel gives A the estate as against O; the recording system gives B protection as a purchaser without notice.

Sabo v. Horvath resolves that conflict in favor of the subsequent purchaser, holding that the earlier grantee's recording, being outside the chain, does not impart constructive notice and does not defeat a later bona fide purchaser. Other jurisdictions treat the after-acquired estate as passing automatically and charge the later purchaser with the earlier record. The practical instruction is identical in both: the grantee whose grantor's title is imperfect should re-record once the grantor's title appears of record.

The shelter rule provides that a person who takes from a purchaser protected by the recording act succeeds to that protection, whether or not the transferee gave value and whether or not the transferee had notice. Its purpose is to preserve the marketability of the protected purchaser's title: if the protected purchaser could not convey the protection, the protection would be worth much less than the statute intends.

In chain analysis the rule is applied at the end, not at the beginning. The examiner first determines whether some person in the chain qualified for protection under the governing statute at the moment of that person's purchase. If so, the protection travels forward with the title, and the interests cut off at that moment remain cut off as to all subsequent takers.

One limitation is uniformly recognized. A person who was himself disqualified — typically because of notice, or because of participation in the wrongful conduct — may not obtain protection by conveying to a protected purchaser and reacquiring the title. The exception prevents the shelter rule from being used to launder a defective claim.

A donee, heir, or devisee of a protected purchaser is ordinarily sheltered, notwithstanding the absence of value, because the protection is an attribute of the title conveyed rather than of the transferee's own status. This is one of the few contexts in which a donee is preferred to an earlier purchaser, and it follows from the structure of the rule rather than from any solicitude for gratuitous takers.

Examination problems rarely involve two claimants. Three and four competing interests are common, and the reliable method is pairwise sequencing. The examiner lists every interest in the order of creation, notes for each the date of creation, the date of recording, whether value was given, and the state of knowledge at the time of payment, and then resolves each adjacent pair under the governing statute before composing the results.

Two rules of composition should be stated explicitly. First, priority determinations are relative, not absolute: it is entirely possible for A to prevail over B, B to prevail over C, and C to prevail over A, producing a circular priority that courts resolve by subrogation, by equitable adjustment, or by treating the middle interest as satisfying the circle. Second, once an interest has been cut off as against a protected purchaser, the shelter rule preserves that result down the chain; the examiner should therefore apply shelter last, after the pairwise analysis is complete.

The examiner should also distinguish the priority of interests from the enforceability of obligations. An easement cut off as against a protected purchaser is unenforceable against that purchaser's land, but the grant remains a contract between the original parties and may support damages. A mortgage subordinated is still a debt. Nothing in the recording system extinguishes personal obligations.

Where mortgages are among the competing interests, this chapter's analysis identifies and sequences them as instruments discovered in examination. The specialized doctrines that govern mortgage priority — purchase-money priority, future advances, replacement and modification, equitable subrogation, and subordination agreements — belong to Chapter 38 and are not reproduced here.

Deeds and mortgages are not the only matters an examination must disclose. A substantial part of professional practice consists of searches conducted outside the deed indexes altogether, in the judgment dockets, tax rolls, mechanics' lien records, court files, and federal filings.

Judgment liens attach by statute, typically upon docketing in the county where the land lies, and typically to all real property then owned or afterward acquired by the debtor in that county. They are searched by the names of the record owners for the statutory period of the lien, and they are renewed or they expire. Because they attach by operation of law rather than by conveyance, they are frequently missed by examiners who search only the grantor index.

Tax liens follow their own statutory logic. General real-property taxes are commonly a first lien on the parcel regardless of recording and regardless of notice, and they must be verified against the collector's records rather than the recorder's. Federal tax liens are filed in designated offices and are searched by name, with their own rules of duration, refiling, and priority against purchasers and security interests.

Mechanics' liens present the sharpest practical trap, because in most States a perfected lien relates back to the commencement of the work or the first delivery of materials, so that a lien filed after closing may take priority from a date before it. Visible construction is therefore both a factual and a legal warning, and the examiner responds with lien waivers, affidavits, indemnities, and, where necessary, an exception in the opinion.

A lis pendens gives record notice of pending litigation affecting title. Its function is to bind subsequent takers to the outcome of the action, and its effect is procedural rather than substantive: the purchaser who takes with a lis pendens of record takes subject to the judgment ultimately entered. Statutes govern its content, its recording, and its cancellation, and the examiner must check both that it exists and that it has been discharged.

Not every link in a chain is a deed between competent individuals dealing for themselves. Title devolves by will and by intestacy, passes through personal representatives, trustees, guardians, conservators, receivers, and bankruptcy trustees, and moves among corporations, partnerships, and limited liability companies by conveyance, merger, dissolution, and conversion. Each such link must be tested for authority.

A devise link requires the will, its admission to probate in a court with jurisdiction, and proof that the parcel passed as the instrument provides; where the decedent owned land in another State, ancillary proceedings are required. An intestacy link requires proof of heirship, which is commonly supplied by decree and supported by affidavit. Both are exposed to omitted heirs, to creditors' claims within the statutory period, and to the time limits on will contests.

A fiduciary link requires the instrument or order creating the office, proof of the fiduciary's qualification and continued authority, and confirmation that the power exercised was granted. A trustee's deed is only as good as the trust's power of sale; a personal representative's deed may require a license or an order of confirmation; a guardian's or conservator's deed almost always requires judicial approval. In many States a certificate of trust or a recorded memorandum supplies the necessary proof without disclosing the whole instrument.

An entity link requires existence, capacity, and authority. The examiner checks good standing, the authority of the signing officer or manager, and any statutory or organic requirement of member, shareholder, or partner approval for a conveyance of substantially all assets. Mergers and conversions transfer title by operation of law in most States, and the certificate is recorded to complete the record chain.

Two survivorship links deserve mention because they are so common. Title held in joint tenancy or by the entirety passes to the survivor outside probate, and the record is completed by a death certificate and affidavit rather than by a deed. Title held in a revocable trust passes according to the trust, and the examiner requires evidence of the successor trustee's appointment.

The recording system generates its own debris. Ancient mortgages never released, tax sales of doubtful regularity, defective acknowledgments, conditions and possibilities of reverter created generations ago, mineral reservations of unknown ownership, and covenants whose beneficiaries have vanished all persist in the record long after they have ceased to matter. Two families of legislation clear it.

Curative acts operate on defects of form. A typical statute provides that an instrument recorded for a stated number of years is not to be denied effect because of an omitted or defective acknowledgment, a missing seal, a failure of attestation, or a similar irregularity. Curative acts validate; they do not extinguish. They are retroactive in operation and are sustained against constitutional objection on the ground that they confirm rather than divest.

Marketable record title acts operate on substance. Modeled on the 1960 Model Marketable Title Act and enacted in roughly a third of the States, they provide that a person with an unbroken record chain from a root of title at least thirty or forty years old holds a marketable record title free of interests arising before that root, unless the interest is preserved by a recorded notice, by an exception in a muniment within the chain, or by an express statutory exception.

The exceptions are the difficulty. Most acts preserve the rights of persons in possession, interests of the United States and often of the State, easements and use rights that are visible or in use, mineral interests in some States, and interests preserved by timely filing. These exceptions are numerous enough that the act is a substantial simplification rather than a complete one, and the criticisms collected in the literature turn largely on their unevenness.

For the examiner the operative instruction is precise. Identify whether the situs has a marketable-title act; identify the root of title; confirm the unbroken chain from the root; search the statutory period for preserving notices; check each exception against the facts; and state in the opinion both the root relied upon and the exceptions taken. Marketable record title is a legal conclusion about the record and is not a substitute for the assurance products taken up in Chapter 35.

The doctrine of this chapter reduces to a method, and the method should be executed in a fixed order. Order matters because each step supplies the inputs of the next, and because an examiner who improvises will omit the steps that seem unnecessary in the ordinary case and are decisive in the extraordinary one.

The method has fourteen steps. Identify the parcel and obtain its legal description from an authoritative source. Identify the governing law of the situs, including the recording statute, the search period, the marketable-title act if any, and local title standards. Determine the starting point and root of title. Search the grantee index backward from the present record owner to the root. Prepare the run sheet. Search each owner forward in the grantor index for the required period. Retrieve and read every instrument disclosed, including plats and referenced documents. Search the non-deed records: judgments, taxes, federal liens, mechanics' liens, lis pendens, bankruptcy, and assessments. Test every non-deed link for authority: probate, fiduciary, entity, and survivorship. Test every link for validity, distinguishing void from voidable defects. Inspect the property, or require an inspection and survey, and pursue every trigger of inquiry notice. Apply the recording statute pairwise to competing interests and compose the result, applying the shelter rule last. Apply curative and marketable-title legislation to the defects that survive. Render the opinion, stating the parcel, the period searched, the owner of record, the estate held, the exceptions taken, and the requirements to be satisfied before closing.

Three professional obligations attach to the last step. The examiner must state what was not searched, because an unstated limitation is functionally a misrepresentation. The examiner must distinguish an exception, which is a matter the opinion does not cover, from a requirement, which is an act to be performed before the opinion becomes reliable. And the examiner must not convert an unresolved doubt into silence: a defect that cannot be cured should be stated as a defect.

The examiner's product is an opinion about the record. It is not a guarantee, it does not reach interests that never appear of record, and it does not insure against the failure of the examination itself. That gap between an accurate examination and an assured title is the institutional problem of Part XII.

The following questions are asked in order of any chain-of-title problem. They are drafted so that a negative answer identifies the doctrine to be applied rather than merely reporting a failure.

1. What is the parcel, and is its description consistent through every instrument in the chain?
2. Which State's recording statute governs, and is it race, notice, or race-notice?
3. What is the root of title, and does a marketable record title act supply or limit it?
4. Is the record chain unbroken from the root to the present owner, and is every link accounted for?
5. For each owner, over what period must the grantor index be run, and was it run for that period?
6. Does any instrument in the chain recite, refer to, or incorporate another instrument, plat, or agreement?
7. Is any recorded instrument outside the chain as wild, early-recorded, late-recorded, or a common-grantor deed?
8. Was every instrument entitled to record, properly acknowledged, and properly indexed?
9. Does any link depend on estoppel by deed or after-acquired title, and has the deed been re-recorded?
10. Is any link void rather than voidable, and if voidable, has a bona fide purchaser intervened?
11. Does any link pass by devise, intestacy, fiduciary act, survivorship, or entity succession, and is the authority proved?
12. What do the judgment, tax, federal lien, mechanics' lien, lis pendens, and bankruptcy searches disclose?
13. Who is in possession, and what would inspection and inquiry of the occupant reveal?
14. Do visible conditions or a survey suggest easements, encroachments, or boundary claims not of record?
15. Applying the statute pairwise, what is the priority of each competing interest, and does the shelter rule alter it?
16. Which defects are cured by curative legislation or extinguished by a marketable-title act, and which survive?
17. What exceptions must the opinion take, and what requirements must be satisfied before closing?
18. What was not searched, and is that limitation stated expressly in the opinion?

Unless otherwise stated, O is the common grantor, A is the earlier claimant, and B is the later purchaser. Each illustration states facts, the governing rule, the analysis, and the result. Where the outcome depends on the recording statute in force, the statute is specified.

1. “If it is recorded, the purchaser is charged with it.” A purchaser is charged only with what a legally sufficient search would reveal. Wild deeds, early and late recordings, indefinite descriptions, and in many States misindexed instruments are recorded and yet impart no constructive notice.
2. “The chain of title is a document you can obtain from the recorder.” It is a reconstruction performed by an examiner from name indexes. The recorder supplies instruments and finding aids, not a chain and not an opinion.
3. “A clean record search means the purchaser takes without notice.” It means only that there is no record notice. Actual knowledge and inquiry notice operate independently, and possession alone will often defeat a purchaser whose search was flawless.
4. “A title search is a search of the deed records.” Judgments, taxes, federal liens, mechanics' liens, lis pendens, bankruptcy filings, probate files, and assessment records are searched separately and are frequently decisive.
5. “Each owner is searched forward only during the period of ownership.” The relevant period is the period of exposure, which ordinarily begins when title vests and ends when the deed out is recorded, and in some States extends further to capture late-recorded instruments.
6. “Estoppel by deed always defeats a later purchaser.” It binds the grantor and passes the after-acquired estate to the earlier grantee, but the earlier recording may be outside the later purchaser's chain, and in jurisdictions following Sabo v. Horvath the later purchaser prevails.
7. “Restrictions in deeds to neighboring lots always bind the retained lot.” That is the Guillette approach. Under Witter v. Taggart such restrictions are outside the retained parcel's direct chain and impart no record notice, though inquiry notice may still reach the purchaser.
8. “Recording an instrument twice serves no purpose.” Re-recording is the standard remedy for the early-recorded deed and for defects cured after the original recording. It places the instrument within the period a searcher will actually run.
9. “A forged deed can be cured by a bona fide purchaser downstream.” A void instrument conveys nothing, and no subsequent purchaser can take title through it. Only voidable instruments yield to a bona fide purchaser.
10. “A quitclaim in the chain makes the title bad.” A quitclaim conveys whatever the grantor has and is common in curative practice. It supplies no covenants and in some jurisdictions bears on bona fide purchaser status, but it is not itself a defect.
11. “A tract index eliminates the need to read instruments.” It improves discovery only. Validity, delivery, authority, description, and meaning are determined from the instruments, whatever index disclosed them.
12. “Electronic and full-text search has enlarged constructive notice.” Constructive notice is measured by the search the law prescribes. Better tools may find more, but they do not by themselves expand the searcher's legal duty.
13. “A marketable record title act guarantees marketable title.” It extinguishes certain pre-root interests unless preserved, subject to substantial exceptions for possession, visible easements, government interests, and timely filings. It clears the record; it does not insure it.
14. “Curative acts and marketable-title acts do the same work.” Curative acts validate instruments defective in form. Marketable-title acts extinguish stale interests measured from a root of title. Their operation, timing, and constitutional justification differ.
15. “A deed from a trustee or personal representative may be accepted at face value.” The fiduciary conveys only within the authority conferred. The examiner requires proof of the office, its continuance, and the specific power exercised.
16. “Mechanics' liens filed after closing cannot affect the purchaser.” In most States a perfected lien relates back to the commencement of work, so visible construction before closing is a warning that no search can answer; waivers and indemnities are the response.
17. “Priority is a single ranking of all interests.” Priority is determined pairwise. Circular priorities are possible and are resolved by equitable adjustment rather than by forcing a linear order.
18. “The examiner's opinion is a guarantee of ownership.” It is a professional judgment about the state of the record for a stated period, subject to stated exceptions and requirements. Assurance against the risks it cannot reach is the subject of Chapter 35.

Chapter 34 has developed the chain of title as the operative unit of American conveyancing practice and has shown that constructive notice is a function of the search the law requires rather than of the contents of the records. The chain is a reconstruction: a skeleton built from name indexes by a backward search and tested by a forward search over each owner's period of exposure, and a substance supplied by reading every instrument the search discloses, together with the plats, agreements, and prior conveyances those instruments recite.

From that definition the chapter derived the categories of recorded instruments that impart no notice — the wild deed, the deed recorded too early, the deed recorded too late, the common-grantor deed to another parcel, the unrecordable instrument, and the instrument whose description identifies nothing — and it examined the jurisdictional divisions those categories produce, from Board of Education v. Hughes and Ayer and Morse to Guillette and Witter, together with the split over recorder indexing errors.

The chapter then addressed what the record cannot close. Inquiry notice arising from possession, visible use, recitals, and irregularity operates independently of the search and defeats purchasers whose examinations were faultless. It distinguished void from voidable links and traced each through the chain, stated estoppel by deed and after-acquired title with their imperfect effect against later purchasers, and applied the shelter rule as the final step of a pairwise priority sequence that may in unusual cases be circular.

Beyond the deed indexes the chapter treated judgments, taxes, federal liens, mechanics' liens with their relation-back priority, lis pendens, and bankruptcy, and the probate, fiduciary, survivorship, and entity links whose validity depends on proof of authority rather than on the form of a deed. It then examined the system's self-repair mechanisms — curative acts, which validate defects of form, and marketable record title acts, which extinguish stale pre-root interests subject to substantial exceptions — and reduced the whole to a fourteen-step examination methodology, an eighteen-question checklist, twenty worked illustrations, and eighteen corrected misconceptions.

Part XI is now complete. Chapter 33 supplied the recording statutes and the doctrine of notice; Chapter 34 supplied the search that gives them content. What neither chapter can supply is certainty. An examination is an opinion about a record, and the record does not disclose forgery, incapacity, undelivered deeds, omitted heirs, unrecorded interests protected by possession, survey defects, or the examiner's own error. Part XII — Title Assurance begins at exactly that limit. Chapter 35 — Title Insurance and Marketable Title takes the results of the examination developed here and asks how the risks it cannot eliminate are allocated: by the marketability standard implied in contracts of sale, by abstract-and-opinion practice, by curative and marketable-title legislation, and by the modern institution of title insurance, with its commitments, exceptions, exclusions, endorsements, and duties to defend.

• Patton and Palomar on Land Titles §§ 12–70 (3d ed.) (search practice, chain construction, and examination standards)
• 4 American Law of Property §§ 17.17–17.29 (A. James Casner ed. 1952) (chain of title and constructive notice)
• 11 Richard R. Powell, Powell on Real Property chs. 82–84, 92
• 4 Herbert Thorndike Tiffany, The Law of Real Property §§ 1259–1318 (3d ed.)
• Paul E. Basye, Clearing Land Titles (2d ed. 1970) (curative legislation and title defects)
• Lewis M. Simes & Clarence B. Taylor, The Improvement of Conveyancing by Legislation (1960) (the Model Marketable Title Act and its rationale)
• Walter E. Barnett, Marketable Title Acts — Panacea or Pandemonium?, 53 Cornell L. Rev. 45 (1967)
• John L. McCormack, Torrens and Recording: Land Title Assurance in the Computer Age, 18 Wm. Mitchell L. Rev. 61 (1992)
• Restatement (Third) of Property: Servitudes §§ 2.14, 7.14 (Am. L. Inst. 2000) (implied servitudes and notice)
• Local title standards promulgated by state bar associations (the authoritative statement of customary search periods and curative practice)