Latent A-10042: Restoring the Forensic Record

Latent A-10042: Restoring the Forensic Record

FBI Vault provenance, ZAP documentary reconstruction, and the CHS cross-consulting system

 

By Space: The archive as evidentiary geography

ZRA9 moves through more than physical crime-scene space. Its principal geography is documentary: the FBI Vault, the reconstructed ZAP archive, and the CHS concordance that cross-consults the two. The location of a document within those systems matters because archival proximity is not necessarily evidentiary identity. A letter latent, a taxi-cab latent, a laboratory comparison, and a later administrative search may occupy the same investigative history without representing the same physical evidence object.

By Time: The long afterlife of A-10042

The relevant chronology extends from the first latent-processing and comparison activity of 1969 through repeated screening in 1969–1970, later FBI reuse, the 1975 material, and the computerized reactivation of A-10042 in 1991. ZRA9 therefore distinguishes historical event chronology from archive storage order. A later FBI page may preserve an earlier evidentiary object; a later search may reuse a decades-old case-set; and administrative continuity across time does not by itself establish continuity of donor identity.

By Motif: Certainty and its boundary

The governing motif is forensic certainty: not whether fingerprint evidence is “good” or “bad,” but what a particular comparison actually establishes. ZRA9 repeatedly encounters the same structural distinction — a definite comparison result may coexist with uncertainty about the provenance of the questioned impression, incomplete anatomical coverage, absent palm or major-case exemplars, or an unresolved elimination history.

The three axes therefore converge on one principle:

A comparison result can be certain while the identity or completeness of the evidence universe remains uncertain.

 

Scope and Position of ZRA9 within the ZRA / CHS Project

ZRA9 is a thematic forensic reconstruction rather than a Section 1–100 documentary walkthrough.

Within the wider ZRA project, its function is methodological as well as historical. Other investigations in the series may begin from an event, a place, a date, a cultural object, or a recurring motif. Here the object of investigation is partly the archive itself: how a forensic proposition survives across evidence submissions, laboratory examinations, suspect comparisons, later administrative reuse, archival rearrangement, and modern reconstruction.

For that reason, ZRA9 also serves as a bridge into the CHS documentary framework.

The article uses selected anchor packets to reconstruct the evidentiary problem surrounding A-10042 while demonstrating the concordance system by which the FBI Vault and the ZAP archive can be cross-consulted without being made artificially identical. The FBI-native page structure, ZAP section and filename system, and CHS analytical identifiers are therefore preserved simultaneously wherever possible.

ZRA9 does not attempt to narrate every ZAP Section from 1 through 100. That complete sequence belongs to a subsequent documentary concordance, where every packet can be examined chronologically and cross-walked against native FBI provenance, request/result relationships, case and specimen identifiers, and unresolved joins.

The present article performs the work necessary before such a walkthrough can be meaningful.

It establishes the rules by which the evidence will be read.

 

A Note on Archive Locators and Crosswalk Status

ZRA9 uses three related but distinct locator systems.

FBI Vault native refers to the surviving FBI release structure: Vault section, PDF segment, and native page location.

ZAP reconstructed refers to the documentary reconstruction used to navigate individual packets through ZAP section numbers and preserved descriptive filenames.

CHS is the project's concordance and analytical layer. It records the relationship between those systems together with source identity, chronology, evidentiary function, comparison state, and uncertainty.

These systems are superimposed for consultation; they are not interchangeable.

Three crosswalk states recur throughout the article:

VERIFIED — the documentary join has been securely established.

PARTIAL — the relationship is established at a broader level, but exact document-to-page mapping remains incomplete.

OPEN — the relevant source or evidentiary branch is known, but the precise native or object-level relationship has not yet been proven.

The presence of an OPEN or PARTIAL locator is therefore not an absence of method. It is the method refusing to manufacture precision.

Where exact page citation and documentary integrity conflict, ZRA9 preserves documentary integrity.

 

Introduction — The Problem of Certainty

Few subjects in the Zodiac case sound more definitive than fingerprints.

A fingerprint either matches or it does not. A latent either identifies a person or it does not. From that apparent simplicity has grown a much larger historical proposition: that investigators possessed a stable body of “Zodiac fingerprints” against which suspects could be conclusively tested.

The FBI record is more complicated.

Across records beginning in 1969 and continuing into the 1991 reactivation of latent case A-10042, the Bureau handled multiple latent fingerprints, palm impressions and other friction-ridge impressions associated with different evidentiary objects and investigative contexts. Some impressions could be compared with ordinary fingerprint records. Others involved fingertips, sides of fingers, lower-joint areas or palm-relevant anatomy for which corresponding exemplars were absent or incomplete. Some impressions were identified to known persons. Others remained unidentified. Some later comparisons produced definite non-identifications while other impressions in the same case-set remained incapable of conclusive comparison with the standards then available.

The distinction is fundamental.

A comparison result can be certain while the identity or completeness of the evidence being compared remains uncertain.

That principle does not weaken a genuine fingerprint exclusion. It defines what the exclusion actually establishes.

ZRA9 therefore does not begin by asking whether a particular suspect “passed” or “failed” the fingerprints. It begins one level earlier:

What evidence existed? Where did it come from? What anatomical area did it represent? What comparison standards were available? What examination was actually performed? What result did the examiner report? And what proposition does that result legitimately support?

The purpose is not to rehabilitate or eliminate a person of interest. It is to restore the chain:

document → evidence object → examination → result → limitation → permissible inference.

 

Key Takeaways

The documentary reconstruction developed in ZRA9 supports six controlling conclusions.

1.     A-10042 is an FBI latent-case history, not a single authenticated “Zodiac fingerprint.”

The records show a long-lived administrative comparison and search case-set containing heterogeneous friction-ridge material. Membership in A-10042 establishes investigative association; offender provenance must still be established impression by impression.

2.     Physical source must remain separate from donor identity.

The October 1969 Stine taxi-cab packet has strong source provenance, and one submitted latent fingerprint was identified as Paul Stine's. Other unidentified cab impressions do not thereby become Zodiac impressions. The separate Stine-letter latent branch remains unresolved until its own FBI disposition is securely recovered.

3.     The 1969–1970 comparison architecture was not simply MATCH / CLEARED.

The FBI repeatedly reported ordinary fingerprint non-identifications while separately confronting absent palms, lower-joint impressions, tips, sides, or other extended-anatomy exemplars. A definite non-match on comparable ordinary fingerprints could therefore coexist with an incomplete comparison of other questioned impressions.

4.     The March 29, 1991 Arthur Leigh Allen result contains both certainty and limitation.

The FBI reported that 38 latent fingerprints were not Allen's fingerprints. The same report separately states that residual tip- and side-area fingerprints and other impressions produced no identification with the fingerprints then available, that no palm prints for the compared individuals were contained in FBI files, and that clearly and completely recorded major-case prints were required for conclusive comparison of the residual material.

The first proposition should not be weakened. The latter propositions should not be erased.

5.     Search failure and administrative membership do not establish offender provenance.

The 1991 automated-search corpus demonstrates extensive searching of A-10042 material. A search that produced no identification did not thereby authenticate the questioned latent as Zodiac's, and the number of ALSA3 subnumbers cannot presently be converted into a number of unique offender fingerprints.

6.     The principal unresolved questions fall into five controlled classes.

They concern provenance and physical-object identity; exemplar completeness and elimination history; numerical integrity; FBI Vault/ZAP/CHS crosswalk completion; and permanent publication controls governing how forensic conclusions are stated.

These findings produce neither a universal rejection nor a universal validation of the Zodiac fingerprint evidence.

They define the evidentiary question more precisely.

 

I. The Object Before the Interpretation

The first mistake to avoid is treating A-10042 as though the designation itself identified a single biological object.

It does not.

The surviving record shows A-10042 operating as a reusable FBI latent-case comparison and search container. By late November and December 1969, Identification Division work sheets and reports explicitly carry A-10042 through repeated suspect comparisons. It remains active into early 1970, appears again in later FBI comparison activity, survives into the 1975 material, and is formally reactivated and computerized in 1991.

Documentary reference [DR-10 / DR-11]. Long-life anchors: FBI Vault Section 4 opening material (August 1970 A-10042 reuse) and FBI Vault Section 5 first segment, zodiac5 pages 1-249(1).pdf, audit clusters pp.3-10, 23-34, and 47-48 for 1975 suspect development. These later records demonstrate institutional persistence of the case-set, not original donor provenance. FBI Vault Section 5 continuation, source zodiac5 pages 250-372.pdf, native pp.259-263; FBI file 9-49911; latent case A-10042; March 29, 1991 Identification Division report/work sheets. This is the controlling packet for the 37/38 fingerprint non-identifications, residual tip/side impressions, absent palms, and major-case-print requirement.

Administrative continuity, however, is not donor continuity.

The fact that a latent was stored or searched as part of a Zodiac-investigation case-set does not establish that the latent was deposited by the offender. That proposition must be established separately for each impression from its physical source, recovery circumstances, elimination history and subsequent examination.

The early record demonstrates why.

In August 1969, the Concerned Citizen submission produced what the reconstructed record describes as a latent palm impression of value on the Q16 envelope. Q17 and Q18 reportedly yielded no latent impressions of value. But the envelope had a handling history. The existence of a useful palm impression establishes that an impression was present on the carrier object; it does not by itself establish that the anonymous writer deposited it.

Documentary reference [DR-01]. CHS/ZFBI §6; ZAP RAW: 1969-08-26 - [6] Examination results P1.jpg and 1969-08-26 - [6] Examination results P2.jpg; FBI case stream 9-49911; Q16-Q18. The ZRA9 native-page join for this packet remains open.

There is an additional caution. The designation “Q16” later reappears in 1991 material in connection with an impression described as potentially originating from a lower joint of a finger or from a palm. The current archive does not assume that the two Q16 references identify the same physical object. FBI Q and Qc designations are case-context identifiers, not numbers that can safely be treated as globally unique across every later submission and reactivation. That crosswalk remains open.

The same discipline applies to Lake Berryessa. The surviving reconstructed record identifies A-12397 as a separate latent-case number associated with that evidence stream. A-12397 and A-10042 therefore remain separate in ZRA9 until a primary administrative document establishes their relationship. Similarity of purpose is not proof of identity.

Documentary reference [DR-02]. CHS/ZFBI §20; ZAP RAW: 1969-10-22 - [20] Request for examination.jpg and 1969-10-29 - [20] Results of examination.jpg; FBI latent case A-12397. Relationship to A-10042 remains unresolved; exact native FBI Vault page join remains open.

The archive thus begins not with one fingerprint but with a set of evidentiary branches whose relationships must be proven rather than presumed.

 

II. What Was Actually Developed

The Paul Stine evidence provides the clearest example of why physical provenance must come before interpretive labels.

There are at least two distinct latent branches in the surviving reconstructed record.

On October 17, 1969, SFPD reported obtaining latent impressions from the October 13 Stine letter and indicated that photographs were to be supplied for FBI searching. ZRA9 has not yet recovered an explicit final FBI disposition tying that branch to a later latent report. It therefore remains open.

Documentary reference [DR-03]. CHS/ZFBI §11; ZAP RAW: 1969-10-17 - [11] Latents found on Stine letter P1.jpg, P2.jpg, 1969-10-20 - [11] Lab Report.jpeg, and the 24 October follow-up 1969-10-24 - [11] Originals not used in comparison.jpg (clean variants are archival duplicates). Final FBI latent disposition and exact native-page join remain open.

Two days later, a different documentary sequence concerns the Paul Stine taxi cab.

The October 19 SFPD memorandum/transmittal identifies the submitted latent photographs as having come from the cab. SFPD further distinguished among those impressions, specially marking blood-associated and right-front-door-handle prints that it believed were more likely to have originated from the suspect. That assessment is historically important, but it remains an investigative assessment rather than a friction-ridge identification of the donor.

Documentary reference [DR-04]. CHS/ZFBI §21; ZAP RAW: 1969-10-19 - [21] SFPD memo to FBI about cab prints.jpg and 1969-10-19 - [21] Request to check Stine cab prints against others.jpg; FBI Vault Section 1, native pp.80-81. Crosswalk verified as the Stine taxi-cab provenance packet.

The October 29 Identification Division material provides the next step. The reconstructed ledger records approximately 30 latent fingerprints, three palm impressions and one lower-joint/palm-type impression within the cab packet. One fingerprint was identified as belonging to Paul Stine. The remaining examined material did not produce identifications under the available comparison basis, with corresponding palm or extended-anatomy standards incomplete for some comparisons.

Documentary reference [DR-05]. CHS/ZFBI §13X; ZAP RAW: 1969-10-29 - [13X] Cover sheet.jpg and 1969-10-29 - [13X] POI and Stine latents P1.jpg through P3.jpg; FBI Vault Section 1, native pp.86-88. Crosswalk verified as the Stine taxi-cab latent result packet.

This result proves something significant.

At least one latent donor in the cab evidence was Stine himself.

It simultaneously warns against another inference: an unidentified impression from the cab does not become a Zodiac impression merely because it remained unidentified.

The cab was a physical environment. Its surfaces could contain impressions from the victim, passengers, mechanics, cleaners, police personnel or other persons who had touched it. SFPD's contemporary judgment that some locations were more probative than others deserves preservation, but investigative plausibility and biometric donor identification are separate propositions.

The Stine-letter branch presents the reverse danger. Because the letter and cab were both associated with the same homicide, a later reader can easily allow their latent histories to merge. The current native crosswalk explicitly prohibits that move: the Stine-letter recovery remains one branch, while the cab memorandum and cab Identification Division result form another.

That distinction may look archival.

It is forensic.

If the source object is wrong, every inference downstream from it is wrong as well.

 

III. How the FBI Used A-10042

By October 1969, A-10042 was already operational as a suspect-screening case. An October 1 request asked the FBI Identification Division to compare a redacted subject against latent fingerprints developed on questioned documents under A-10042; the October 6 result reported that the previously reported latent fingerprints were not identical with the tested individual's fingerprints while separately noting that no palm prints for that individual were contained in FBI files. By November, the record becomes even more revealing because it shows what the Bureau considered necessary for conclusive comparison beyond ordinary rolled fingerprints.

Documentary reference [DR-06]. CHS/ZFBI §§8-9; ZAP RAW §8: 1969-10-01 - [08] Compare poi latents P1.jpg, P2.jpg (plus typed transcription); ZAP RAW §9: 1969-10-01 - [9] Radiogram request for examination.jpg and 1969-10-06 - [9] Result of examination P1.jpg through P3.jpg; latent case A-10042. This request/result pair documents an operational A-10042 suspect comparison and absence of palm prints in FBI files for the tested individual.

One A-10042 comparison recorded an ordinary fingerprint non-identification while separately requiring impressions from the extreme tips, sides and lower joints of the fingers, together with palms, for conclusive comparison of other questioned material.

Documentary reference [DR-07]. CHS/ZFBI §28; ZAP RAW: 1969-11-09 - [28] Result of examination.jpg; A-10042 stream. The result preserves an ordinary-print non-identification together with a request for extreme tips, sides, lower joints, and palms for conclusive comparison. Exact native FBI Vault page join remains open.

That is not merely a technical qualification attached decades later.

It is part of the original comparison architecture.

The distinction recurs.

During late November and December 1969, A-10042 was used in a broad multi-subject screening program. Identification Division work sheets and reports preserve repeated comparisons against named or subsequently redacted individuals. In several reconstructed sequences, available ordinary fingerprints produced non-identifications while corresponding lower-joint or palm exemplars remained unavailable.

The December material is particularly useful because the same limitation appears across different comparison packets.

The veterans-hospital cohort represented in Section 59 produced ordinary-print non-identifications while an ambiguous lower-joint/palm-type impression could not be completely resolved because the corresponding standards were absent.

Section 63 records another individual whose available ordinary fingerprints did not identify him as the source, while the relevant lower-joint/palm counterpart remained unavailable.

Section 64 expands the process to a multi-person roster. Different subjects had different record-availability states, and the lower-joint/palm problem recurred. The surviving release is too redacted and degraded to justify manufacturing a complete person-by-person disposition table.

Documentary reference [DR-08]. CHS/ZFBI §§59, 63, 64, and 71; ZRA9 comparison-history records P007-P010; FBI Vault Section 3, native pp.4-19 at cluster level. The relationship is one-to-many and remains partial until each ZAP packet is frozen to exact native pages.

The FBI was therefore not operating with a binary system of:

MATCH / CLEARED.

The archive instead preserves several distinct states:

No usable identification record located.

Available ordinary fingerprints compared and found non-identical.

No identification obtained from an impression with available standards.

Corresponding palm/lower-joint/tip/side exemplar unavailable.

Additional major-case impressions required for conclusive comparison.

Those states are not synonyms.

Early 1970 comparison cycles continue the same pattern. Some surviving packets are degraded; some contain a result without a securely reconstructed request; some establish only that no positive identification occurred. ZRA9 therefore refuses to reconstruct stronger dispositions from redactions or incomplete paperwork.

Documentary reference [DR-09]. CHS/ZFBI §§85, 86, 87, 92, and 93; ZRA9 comparison-history P011-P014. Section 85 is RESULT-ONLY / DEGRADED; Sections 86-87 and 92-93 preserve continuing A-10042 request/result chains. Exact native joins remain partial and redacted identities are not reconstructed.

And A-10042 did not disappear when the initial Zodiac investigation receded from its peak.

The FBI Vault chronology shows renewed comparison activity in August 1970. Later Section 5 material shows the case-set still being used for suspect screening in 1975, with record-availability and exemplar-completeness problems recurring years after its creation.

Documentary reference [DR-10]. FBI Vault Section 4 opening material (August 1970 A-10042 reuse) and FBI Vault Section 5 first segment, zodiac5 pages 1-249(1).pdf, audit clusters pp.3-10, 23-34, and 47-48 for 1975 suspect development. These later records demonstrate institutional persistence of the case-set, not original donor provenance.

That longevity changes how A-10042 should be understood.

It was not simply a snapshot of one crime scene.

It became an operational repository against which new people could be screened over time.

 

IV. “Not Identical” Is a Result, Not a Biography

The distinction becomes unavoidable in the 1991 reactivation.

On March 29, 1991, the FBI Identification Division issued the most useful surviving statement presently audited for understanding the structure of A-10042.

The report separated the evidence into different comparison states.

For one redacted subject, 37 previously reported latent fingerprints were reported not to be that person's fingerprints.

For Arthur Leigh Allen, 38 latent fingerprints were reported not to be his fingerprints.

Those are definite fingerprint non-identifications. They should not be weakened, hedged away or reinterpreted as though the FBI did not reach them.

But the report did not stop there.

It separately addressed the remaining latent fingerprints from the tip and side areas of the fingers and other latent impressions. Those materials were compared with the fingerprints then available, and no identification was effected.

The same report stated that no palm prints were contained in FBI files for the compared individuals.

And it further stated that clearly and completely recorded major-case prints were required for conclusive comparison of the residual material.

Documentary reference [DR-11]. FBI Vault Section 5 continuation, source zodiac5 pages 250-372.pdf, native pp.259-263; FBI file 9-49911; latent case A-10042; March 29, 1991 Identification Division report/work sheets. This is the controlling packet for the 37/38 fingerprint non-identifications, residual tip/side impressions, absent palms, and major-case-print requirement.

These statements describe different propositions.

For the 38 comparable fingerprints:

Allen was not the donor of those 38 fingerprints.

For the residual impressions:

the available exemplars did not produce an identification.

For the absent palm standards:

a palm comparison could not be completed from palms that the FBI did not possess.

For the requested major-case prints:

the existing exemplar set was insufficient for conclusive comparison of all relevant anatomy.

It therefore exceeds the audited packet to rewrite the March 1991 report as either:

“Allen matched the Zodiac fingerprints.”

or:

“Allen was conclusively eliminated by every relevant Zodiac fingerprint and palm impression.”

Neither formulation represents the structure of the document.

The report instead preserves something considerably more useful: definite non-identification and incomplete comparison can coexist within the same case, the same suspect and the same report.

This distinction is not special pleading for Allen. The same reasoning must be applied to every person in the archive.

When ordinary fingerprints were actually compared and found non-identical, that result is real.

When palms were unavailable, that absence is also real.

When no Bureau identification record could be located for a person, that is a records-search result—not a fingerprint comparison.

When an impression could not be conclusively compared because the necessary anatomy was missing from the exemplar set, that limitation cannot be converted into either a match or an exclusion.

ZRA9 therefore treats a forensic result as a proposition about the material actually compared, not a biography of the person who supplied the exemplar.

 

V. The Long Afterlife of A-10042

The 1991 reactivation also reveals how the Bureau itself conceptualized A-10042 after more than two decades.

Sacramento requested photographs of all latent fingerprints developed during the Zodiac investigation under A-10042 and sought automated searches of the material. Vallejo intended to use automated systems as well.

The same request sequence specifically noted that at least one impression on a prior Q16 submission came from a lower-joint finger area or a palm area, and broader major-case exemplars were being pursued precisely because ordinary file fingerprints did not represent all necessary anatomy.

Documentary reference [DR-12]. FBI Vault Section 5 continuation, zodiac5 pages 250-372.pdf, native pp.264-268; January 14-23, 1991 Sacramento/Vallejo requests. The packet requests photographs/searches of A-10042 material and notes a prior Q16 impression as lower-joint or palm-relevant. Object identity of that Q16 remains crosswalk-open.

The phrase “all latent fingerprints developed during the Zodiac investigation” is important.

But it is an administrative description of the case-set.

It is not the same statement as:

“all fingerprints in A-10042 were deposited by Zodiac.”

The latter would require donor attribution that the present archive does not provide impression by impression.

The automated-search records add another layer.

During February and March 1991, a large ALSA3 corpus was generated under numerous A-10042 subnumbers. These documents demonstrate sustained computerized search activity. Where a search produced no identification, the proper conclusion is that the automated system did not identify the searched latent under the search conditions represented by that record.

A failed database search does not authenticate the unknown donor.

Nor can the number of ALSA3 subnumbers presently be used as a count of unique “Zodiac fingerprints.” Subnumbers may represent latent records, encodings, searches, duplicates, consolidations or administrative processing. A record-level crosswalk is required before a unique-latent total can responsibly be stated.

Documentary reference [DR-13]. FBI Vault Section 5 continuation, zodiac5 pages 250-372.pdf, native pp.269-366; ALSA3 A-10042.* automated coding/search corpus. Subnumbers are search/information-system records and are not treated as unique offender-print counts.

The distinction matters because numbers acquire authority very quickly in historical forensic discussion.

“Thirty-eight,” for example, is real in the 1991 Allen comparison.

But even there, a second numerical question remains unresolved: whether the report's 37 latents for the redacted suspect and 38 for Allen represent the same underlying core set with a difference in comparability, overlapping subsets, or different sets.

The archive does not yet permit that question to be answered.

So it remains unanswered.

 

VI. The CHS Cross-Consulting System: FBI Vault, ZAP, and Project Provenance

ZRA9 depends on three documentary layers that answer different questions. They must be consulted together, but they must never be collapsed into one numbering system.

CHS is not a third primary source. It is the project concordance layer above two source architectures: the FBI Vault and the ZAP reconstructed archive. CHS employs an audit protocol that populates that concordance while preserving the original labels below it.

 

Table 1. The three documentary layers used by ZRA9

Layer	Primary function	Typical locator	Control rule
FBI Vault native	Preserves Bureau storage, section/page order, file numbers, FOIA gaps, and later institutional reuse.	Vault 1 pp.80-81; Vault 5 pp.259-263	Never renumber native pages to imitate ZAP/CHS.
ZAP reconstructed archive	Preserves reconstructed packet sequence and direct image filenames used to navigate released documents chronologically.	1969-10-19 - [21] SFPD memo to FBI about cab prints.jpg	A ZAP section or filename is a working locator, not an FBI-native section number.
CHS	Cross-consults both systems and records object identity, request/result status, evidentiary lane, correspondence state, and uncertainty.	ZRA9 X003 / L002 / P018; ZFBI §21	Project IDs sit above native labels; they do not replace them.

 

 

VI.A. What a CHS record stores

A CHS concordance record is a composite locator. At minimum, it preserves the project record ID, ZAP section and original filename, FBI Vault section and native page where known, FBI file or latent-case identifier, Q/Qc designation where relevant, document function, request/result state, crosswalk status, and an evidentiary control statement. If one of those joins is not proven, the record says OPEN or PARTIAL instead of forcing a match.

The present ZRA9 data spine already implements this at several linked levels: S### records sources; L### records friction-ridge evidence; C### records chronology; X### records FBI-Vault/ZAP joins; P### records comparisons; CL### records publication claims; and G### records the logic boundary between source statement, permissible inference, and impermissible escalation.

 

Table 2. ZRA9 CHS identifier families

Identifier	Layer	Question answered
S001-S020	Source index	What documentary object or packet is being cited, in which source layer and case stream?
L001-L016	Latent ledger	What friction-ridge object/anatomy/result/limitation is documented?
C001-C023	Chronology	When did the request, examination, result, or reuse occur independently of FBI Vault storage order?
X001-X012	Native crosswalk	How does a ZAP/CHS locator relate to an FBI Vault section/page, and is the join verified, partial, or open?
P001-P019	Comparison history	Who or what was compared, with which exemplar classes available, and what controlled disposition followed?
CL001-CL012	Claim matrix	How strong is a proposed publication claim, and what safer formulation does the evidence support?
G001-G012	Logic audit	What inference is permitted by the source statement, and what escalation would exceed it?

 

VI.B. One object can have several valid locators

The Stine taxi-cab packet demonstrates the method. In the reconstructed ZAP sequence, Section 21 preserves the October 19 cab-provenance documents and Section 13X preserves the October 29 latent result. In the FBI Vault, the same evidentiary chain appears at Section 1 native pp.80-81 and pp.86-88. CHS records both locators and joins them at the object/function level. It does not rename FBI Vault p.80 as “Section 21,” and it does not treat the ZAP section number as an FBI-native archival fact.

Documentary reference [DR-04 / DR-05]. CHS/ZFBI §21; ZAP RAW: 1969-10-19 - [21] SFPD memo to FBI about cab prints.jpg and 1969-10-19 - [21] Request to check Stine cab prints against others.jpg; FBI Vault Section 1, native pp.80-81. Crosswalk verified as the Stine taxi-cab provenance packet. CHS/ZFBI §13X; ZAP RAW: 1969-10-29 - [13X] Cover sheet.jpg and 1969-10-29 - [13X] POI and Stine latents P1.jpg through P3.jpg; FBI Vault Section 1, native pp.86-88. Crosswalk verified as the Stine taxi-cab latent result packet.

 

VI.C. One number can refer to different objects

The reverse problem also occurs. FBI Q/Qc numbers are meaningful only inside their document and case context. Section 81 preserves a “Q1 demand note” belonging to a separate bank-robbery comparison. That Q1 cannot be conflated with Zodiac-file Q1. The case/file context therefore forms part of the specimen identity in CHS.

Documentary reference [DR-14]. CHS method control example, ZFBI §81: ZAP RAW includes 1970-01-08 - [81] Request for examination(1).jpg and the January 16 result pages (with later filename variants dated January 18). The packet's Q1 demand note belongs to a separate bank-robbery case context, demonstrating why Q/Qc labels cannot be treated as globally unique without their case/file context.

 

VI.D. Why the complete Sections 1-100 walkthrough is a follow-up

This article uses selected documentary nodes because its purpose is analytical: to establish how certainty, provenance, anatomy, and exemplar completeness interact inside A-10042. A full Sections 1-100 walkthrough is a different scholarly object - a documentary concordance. The CHS architecture introduced here is the mechanism that will allow that later article to proceed section by section without losing the FBI-native page structure, ZAP filenames, request/result chains, or open joins.

In that sense, ZRA9 does set the stage for the later walkthrough. But it does more than announce it: the tables and documentary references below demonstrate the cross-consulting system already in operation.

 

VII. Archive Versus Description

ZRA9 exists because the Zodiac record has acquired more than one documentary structure.

The FBI Vault preserves how the Bureau stored, grouped, transmitted, revisited and operationally reused material.

The ZAP reconstruction and the CHS layer serve different functions. The ZAP archive preserves a reconstructed packet sequence and direct file-name system that makes fragmented releases navigable. CHS cross-consults that reconstruction against FBI Vault provenance, preserves both locator systems, and records the evidentiary relationship between them.

Those structures should be crosswalked.

They should not be made artificially identical.

The native crosswalk already demonstrates one-to-many and many-to-one relationships. Late-1969 A-10042 screening in FBI Vault 3 spans several reconstructed sections. Early-1970 reconstructed sections cross native boundaries imperfectly. FBI Vault Section 5 itself later extends across an artificial upload break between page 249 and page 250 while remaining a single native FBI section.

Documentary reference [CHS crosswalk]. See ZRA9_NATIVE_CROSSWALK.csv, especially X002-X008 and X011-X012. Verified joins are distinguished from PARTIAL and OPEN joins; the p.249 -> p.250 boundary is an upload split, not a new FBI native section.

The same rule applies to descriptive certainty.

A later heading or summary may make a complex record easier to understand, but the heading is not the underlying evidence.

For that reason ZRA9 maintains separate conceptual fields rather than a single blended description:

ZAP / secondary description

FBI source statement

CHS project evidentiary assessment

The current audit has already corrected one consequential example by separating the Stine-letter latent branch from the Stine-cab latent branch. The cab result must not silently migrate into the letter history merely because both involve Stine.

There is an important limitation here as well.

The present ZRA9_ZAP_DESCRIPTION_AUDIT.csv is based on CHS/WERK-FBI project descriptions and reconstructed section states. The original ZAP headings and metadata have not yet been separately extracted and frozen verbatim. ZRA9 therefore should not attribute the present audit labels to ZAP contributors as quotations or claims until that metadata extraction is completed.

Documentary reference [ZAP metadata control]. See ZRA9_ZAP_DESCRIPTION_AUDIT.csv. Its current descriptions are CHS/WERK-FBI project labels and reconstructed section states, not claimed verbatim ZAP headings; original ZAP metadata extraction remains an open archival task.

That distinction is not merely diplomatic.

It is the same evidentiary discipline being applied to our own archive.

We cannot demand precision from an FBI document and then become imprecise when describing the provenance of a modern archival label.

 

VIII. What Can Be Said With Confidence

The reconstructed record presently supports several findings strongly.

First: A-10042 functioned as a long-lived FBI latent comparison and search case-set associated with the Zodiac investigation. It was used repeatedly from late 1969 through later screening and the 1991 reactivation.

Second: the material cannot responsibly be reduced to a single homogeneous “Zodiac fingerprint.” The records distinguish ordinary fingerprints, palm impressions, tip and side areas, lower-joint/palm-relevant material and automated search records.

Third: the Stine cab has strong physical-source provenance in the October 1969 FBI/SFPD sequence. One submitted cab fingerprint was identified as Paul Stine's. Other unidentified impressions in that packet are not thereby identified as the killer's.

Fourth: the FBI repeatedly distinguished ordinary fingerprint non-identification from complete anatomical comparison. Requests for tips, sides, lower joints, palms and later major-case prints show that the limitation was embedded in the historical record rather than invented retrospectively.

Fifth: in 1991, 38 compared latent fingerprints did not identify Arthur Leigh Allen. The same report separately records residual impressions for which available exemplars did not produce an identification, missing palm standards, and a need for major-case prints before conclusive comparison.

Sixth: neither an automated-search failure nor membership in the A-10042 case-set establishes offender provenance. The donor question remains impression-specific.

These findings do not reduce the value of the fingerprint evidence.

They define it more accurately.

 

 

IX. What Remains Unknown

The archive is now coherent enough to show precisely where its gaps remain.

We still need the final FBI disposition, if one survives, for the latent photographs reportedly obtained from the October 13 Stine letter.

We still need to establish the physical objects behind the 1991 Q2, Q3 and prior Q16 references rather than relying on number similarity.

We still need a document-level map showing which A-10042 entries derive from the Stine cab, correspondence, other questioned documents, or other physical evidence.

We do not yet know the administrative relationship between A-10042 and A-12397.

We have not yet recovered a complete elimination-print universe for victims, police, postal or newspaper handlers, laboratory personnel, cab personnel or other known potential donors.

We do not yet know whether major-case and palm exemplars were subsequently obtained for Allen or other 1991 subjects and, if so, whether a later conclusive residual comparison was performed.

We cannot yet safely state whether the 37- and 38-latent groups in the March 1991 report are identical, overlapping or different subsets.

And we cannot convert ALSA3 subnumbers into a unique latent count until the search, duplication and consolidation relationships are reconstructed.

These are not weaknesses to conceal.

They are part of the result.

A serious forensic reconstruction should identify the point at which the surviving evidence stops.

 

Documentary Roadmap Beyond ZRA9

The next logical companion is a complete CHS concordance of ZAP Sections 1-100 against the native FBI Vault. That work can be chronological, documentary, or both, but its basic unit should be the crosswalk record rather than the prose paragraph. Each section should preserve the original ZAP filename(s), CHS/WERK-FBI state, FBI Vault section/page, case and specimen identifiers, request/result relationship, correspondence status, and unresolved joins. ZRA9 provides the forensic thesis and the indexing grammar; the later concordance can provide the exhaustive walk-through.

 

Conclusion — Restoring the Chain

The fingerprint question in the Zodiac case has often been presented as though only two possibilities exist.

Either investigators possessed genuine Zodiac fingerprints, in which case suspects can be definitively cleared by them;

or the fingerprint evidence is so uncertain that the comparisons themselves have little value.

The FBI record supports neither extreme.

A latent fingerprint comparison can produce a definitive result about the two impressions actually compared. If an examiner determines that a questioned fingerprint and a suspect's fingerprint are not identical, that result is meaningful and should be stated plainly.

But it answers a bounded question.

It does not automatically establish who deposited the questioned impression.

It does not automatically establish that the questioned impression belonged to the offender.

It does not establish that all relevant impressions came from one donor.

And it does not establish that every anatomical area represented in the questioned evidence was also represented in the suspect's available exemplars.

The history of A-10042 repeatedly preserves those distinctions.

In 1969, ordinary fingerprint non-identifications could coexist with requests for fingertips, sides, lower joints and palms.

In subsequent years the same latent case continued to be used as an operational screening repository.

And in 1991 the distinction became explicit again: 38 fingerprints did not identify Arthur Leigh Allen, while other residual impressions remained subject to incomplete exemplar coverage, absent palms and the need for major-case prints before conclusive comparison.

The apparent contradiction disappears once the propositions are separated.

A comparison result can be certain while the identity or completeness of the evidence universe remains uncertain.

That is not an argument for a suspect.

It is not an argument against a suspect.

It is an argument for reading the forensic record at the level at which it was actually created.

And that is the purpose of ZRA9:

not to make the archive say less than it says,

not to make it say more,

but to restore the chain between object, examination, result, limitation and conclusion.

 

SDS

Drafted: 16 August 2026

97 years since the birthday of Norma Jean Hightower Suennen

Published: 20 August 2026

 

 

 

 

Appendix A — Selected CHS Concordance and Direct Documentary References

This table is deliberately selective. It demonstrates the CHS cross-consulting method with the anchor packets used in ZRA9. OPEN and PARTIAL are positive archival states: they record where an exact native or object-level join has not yet been proven.

Table 3. Selected FBI Vault ↔ ZAP ↔ CHS concordance

Ref	CHS / event	ZAP RAW / source file	FBI native locator	FBI ID	Crosswalk / control
DR-01	§6 / Concerned Citizen palm	1969-08-26 - [6] Examination results P1.jpg; P2.jpg	Vault 1 native page join OPEN	9-49911; Q16-Q18	Object/result verified in ZAP/CHS; donor unresolved.
DR-06	§§8-9 / first operational A-10042 screen	§8: 1969-10-01 - [08] Compare poi latents P1/P2.jpg; §9: 1969-10-01 radiogram + 1969-10-06 result P1-P3.jpg	Exact native page join OPEN	A-10042	Request/result verified; fingerprints non-identical; palms absent in FBI file.
DR-03	§11 / Stine-letter latent branch	1969-10-17 - [11] Latents found on Stine letter P1/P2.jpg; 1969-10-20 Lab Report.jpeg; 1969-10-24 follow-up	Vault 1 exact page join OPEN	9-49911; Qc19/Qc20 context	Separate from cab branch; final latent disposition OPEN.
DR-04	§21 / Stine cab provenance	1969-10-19 - [21] SFPD memo to FBI about cab prints.jpg; Request to check Stine cab prints against others.jpg	Vault 1 pp.80-81	9-49911	VERIFIED same cab-provenance packet; agency belief ≠ donor ID.
DR-05	§13X / Stine cab result	1969-10-29 - [13X] Cover sheet.jpg; POI and Stine latents P1-P3.jpg	Vault 1 pp.86-88	9-49911	VERIFIED cab-result packet; one fingerprint Stine; unidentified ≠ Zodiac.
DR-02	§20 / Lake Berryessa	1969-10-22 - [20] Request for examination.jpg; 1969-10-29 - [20] Results of examination.jpg	Exact native page join OPEN	A-12397	Keep separate from A-10042 pending administrative proof.
DR-07	§28 / expanded anatomical exemplars	1969-11-09 - [28] Result of examination.jpg	Vault 1/2 exact page join OPEN	A-10042 stream	Ordinary non-match plus tips/sides/lower joints/palms needed.
DR-08	§§59/63/64/71 / Dec. screening	Original RAW packets preserved by ZAP section; direct file-by-file native join still pending	Vault 3 pp.4-19 cluster	A-10042	PARTIAL one-to-many crosswalk; recurrent lower-joint/palm limits.
DR-09	§§85/86/87/92/93 / early 1970	Original RAW packets preserved by ZAP section	Vault 3 tail / transition; exact joins PARTIAL	A-10042	Result-only/degraded and request/result states kept separate.
DR-14	§81 / Q-number control example	1970-01-08 - [81] Request for examination(1).jpg; January 16 result pages	Separate bank-robbery case context	Q1 demand note	Method control: Q/Qc numbers are not globally unique.
DR-11	1991 A-10042 report	zodiac5 pages 250-372.pdf	Vault 5 pp.259-263	9-49911; A-10042	VERIFIED native; 37/38 definite fingerprint non-IDs + residual incomplete comparison.
DR-12	1991 Sacramento/Vallejo requests	zodiac5 pages 250-372.pdf	Vault 5 pp.264-268	A-10042; Q2/Q3; prior Q16	VERIFIED request; prior Q16 object identity OPEN.
DR-13	1991 ALSA3 search corpus	zodiac5 pages 250-372.pdf	Vault 5 pp.269-366	A-10042.*	VERIFIED corpus; subnumbers ≠ unique offender-print count.

 

 

APPENDIX B — THE CHS ZRA9 DATA SPINE AND RESEARCH APPARATUS

B.1 Purpose of the Data Spine

ZRA9 is supported by a structured research apparatus designed to preserve the distinction between a historical document, the evidence object described by that document, the examination performed, the reported result, the limitations of that result, and the inference that may legitimately be drawn from it.

The governing sequence is:

document → evidence object → examination → result → limitation → permissible inference

The data spine does not function as a substitute for the FBI record. Nor does it replace the ZAP archive or its reconstruction of the documentary sequence. Its purpose is to create a third, auditable layer through which the two source structures can be cross-consulted without forcing them into artificial equivalence.

The FBI Vault preserves the Bureau's surviving documentary organization: native file divisions, page sequence, administrative grouping, later reuse, and the order in which material was released. The ZAP archive preserves another indispensable layer: reconstructed documentary clusters, research sections, descriptive filenames, and a chronology that often makes fragmented FBI material substantially easier to follow.

The CHS system treats those structures as complementary but non-identical.

A ZAP section is therefore not automatically assumed to equal one FBI Vault page range. A single reconstructed section may correspond to several native pages; several reconstructed sections may belong to one FBI administrative packet; and some later native FBI material has no equivalent within the ZAP Sections 1–100 sequence presently being used. Where the relationship is proven, the CHS crosswalk records it as verified. Where the relationship is only partly reconstructed, it remains partial. Where the join has not yet been established, it remains open.

The data spine was frozen in its present form on August 16, 2026, as DATA SPINE v1, reconciled through FBI Vault Section 5, native page 372.

It is composed of two state documents, eight principal analytical CSV tables, a consolidated spreadsheet workbook, and a cryptographic package manifest.

The apparatus is deliberately modular. No single table is intended to carry the whole argument. Instead, each file answers a different evidentiary question.

 

B.2 ZRA9_MASTER.md — Controlling State and Methodological Charter

Function:
The master state file defines the purpose, methodological boundaries, governing principle, current evidentiary conclusions, data-spine architecture, and publication guardrails of ZRA9.

It records the central proposition upon which the article is constructed:

A comparison result can be certain while the identity or completeness of the evidence being compared remains uncertain.

The master file also establishes several controlling findings.

A-10042 operated as a long-lived FBI administrative latent-case comparison and search container. The evidence associated with that case-set was heterogeneous. Ordinary fingerprints, tip and side impressions, lower-joint or palm-relevant material, and later automated-search records cannot responsibly be collapsed into one homogeneous category called “the Zodiac fingerprints.”

The master file further preserves the separation between the Stine taxi-cab latent packet and the still-unresolved Stine-letter latent branch; maintains Lake Berryessa case A-12397 separately from A-10042 pending documentary proof of their administrative relationship; and records the recurrent historical distinction between ordinary fingerprint non-identification and incomplete comparison of additional friction-ridge anatomy.

Its publication guardrail is equally important. ZRA9 does not state that the FBI possessed a single definitive set of authenticated “Zodiac fingerprints.” The historically supportable formulation is narrower: the FBI maintained and repeatedly searched latent evidence associated administratively with the Zodiac investigation, while the provenance, anatomical type, exemplar coverage, and elimination status of individual impressions varied.

CHS citation form: ZRA9-MASTER

 

B.3 ZRA9_SOURCE_INDEX.csv — Documentary Provenance Index

Function:
The Source Index is the principal documentary inventory. Its unit of analysis is a source packet, document sequence, or reconstructed evidentiary cluster rather than an individual inference.

Its fields are:

source_id
source_layer
source_file_or_section
native_vault
native_pages_or_zap_section
event_date
date_precision
case_id
q_designations
evidence_lane
record_function
document_status
provenance_note

The table allows a source to be located simultaneously within its documentary layer and within the chronology of the investigation.

The source_layer field distinguishes FBI-native material from ZAP/CHS reconstruction and from records for which both structures have been securely joined. The document_status field prevents an unresolved or degraded source from acquiring the same apparent authority as a fully verified packet. The provenance_note field records the interpretive control that must accompany use of the source.

Representative examples include:

S004 — Stine-letter latent branch.

This record preserves the October 17, 1969 report that SFPD had obtained latent impressions from the October 13 Stine letter and intended to supply photographs for FBI searching. The branch is verified as a request or planned-search sequence, but its final FBI disposition remains open. The Source Index therefore prohibits silently substituting the later taxi-cab result for the letter-latent branch.

S005 — Stine taxi-cab provenance.

This joins FBI Vault 1, pp.80–81, with ZAP Section 21. It establishes that the latent photographs in that submission came from Paul Stine's taxi cab and separately preserves SFPD's contemporary judgment concerning which prints it believed were more likely to be suspect-origin.

S006 — Stine taxi-cab result.

This joins FBI Vault 1, pp.86–88, with the corrected Section 13X packet. Approximately thirty latent fingerprints, three palm impressions, and one lower-joint/palm-type impression are represented in the reconstructed packet. One fingerprint was identified to Paul Stine. The record explicitly prevents the further inference that the unidentified impressions were therefore deposited by Zodiac.

S017–S019 — 1991 reactivation.

These records preserve the March 1991 Identification Division report, the January reactivation requests, and the large ALSA3 automated-search corpus while preventing administrative case membership or search-record counts from being converted into claims of offender provenance.

The Source Index therefore answers the question:

Where does this claim come from, and what kind of source is it?

CHS citation form: ZRA9-SOURCE S###

Example: ZRA9-SOURCE S006

 

B.4 ZRA9_A10042_LATENT_LEDGER.csv — Core Friction-Ridge Evidence Ledger

Function:
The Latent Ledger is the principal evidence-state table for ZRA9. Whereas the Source Index is document-centered, the Latent Ledger is evidence-centered.

Its fields are:

ledger_id
event_date
date_precision
originating_evidence_or_context
fbi_designation
impression_type
anatomical_classification
quality_or_value_statement
comparison_subject_or_set
palm_availability
major_case_availability
result
source_language_or_summary
establishes
does_not_establish
source_ref
confidence

The critical feature is the simultaneous presence of the fields establishes and does_not_establish.

The ledger is not allowed merely to record a result. It must also state the boundary of that result.

For example:

L002 — Stine taxi cab, October 19, 1969.

The physical source of the submitted latent photographs is established as the cab. SFPD's contemporary prioritization of blood-associated and right-front-door-handle impressions is preserved. What is not established is that every latent — including every specially marked latent — was actually deposited by the offender.

L003 — Stine taxi-cab result, October 29, 1969.

One fingerprint was identified as Paul Stine's. Other submitted impressions remained unidentified or nonidentified under the available comparison basis, with reference palms or extended-anatomy standards incomplete in some instances. The permissible conclusion is that at least one donor in the cab packet was Stine and that additional unidentified impressions existed. The impermissible conclusion is that unidentified equals Zodiac.

L006 — Expanded anatomical comparison requirement, November 1969.

Comparable ordinary fingerprints were non-identical, while extreme tips, sides, lower joints, and palms were required for conclusive comparison of other questioned impressions. The ledger therefore records both a definite result and an incomplete comparison universe.

L013 — Arthur Leigh Allen, March 29, 1991.

Thirty-eight previously reported latent fingerprints were reported not to be Allen's fingerprints. This is recorded as a definite non-identification for those thirty-eight comparable fingerprints.

L014 — Residual 1991 material.

Remaining tip- and side-area latent fingerprints and other impressions produced no identification with the available fingerprints. No palm prints were in FBI files for the compared individuals, and clearly and completely recorded major-case prints were required for conclusive comparison. This evidence state is deliberately maintained separately from L013.

The Latent Ledger therefore answers:

What was the evidence, what anatomical material did it represent, what was actually compared, and what did the result establish — and not establish?

CHS citation form: ZRA9-LATENT L###

Example: ZRA9-LATENT L013

 

B.5 ZRA9_CHRONOLOGY.csv — Event Chronology Independent of Archive Order

Function:
The Chronology reconstructs investigative and forensic events according to historical date rather than FBI Vault storage order or ZAP section number.

Its fields are:

chron_id
event_date
date_precision
event
source_layer
source_locator
case_id
evidentiary_state
control_note

The date_precision field distinguishes exact dates from ranges, months, or years so that uncertain chronology is not converted into false precision.

The evidentiary_state field classifies what kind of historical event is represented: request, intake, recovery, laboratory processing, latent result, non-identification, incomplete comparison, automated search, administrative continuation, authentication review, or another defined state.

The chronology begins with the initial August 1969 federal submissions, passes through the Concerned Citizen material, the Stine-letter and Stine-cab branches, Lake Berryessa A-12397, the creation and repeated use of A-10042, the 1970 and 1975 comparison cycles, separate 1978 authentication and cryptanalytic activity, later external Zodiac-like submissions, and finally the 1991 reactivation.

One of its most important entries is C022, dated March 29, 1991. It records the simultaneous existence of thirty-seven fingerprint non-identifications for one redacted subject, thirty-eight fingerprint non-identifications for Arthur Leigh Allen, and residual material whose comparison basis remained incomplete.

The chronology is intentionally independent of archival storage order because FBI filing order, FOIA release order, and historical event order are not the same thing.

It therefore answers:

When did the evidentiary event actually occur?

CHS citation form: ZRA9-CHRON C###

Example: ZRA9-CHRON C022

 

B.6 ZRA9_NATIVE_CROSSWALK.csv — FBI Vault / ZAP / CHS Concordance

Function:
The Native Crosswalk records the relationship between the FBI Vault's native organization and the ZAP/CHS reconstructed section system.

Its fields are:

cross_id
event_or_object
native_fbi_locator
zap_chs_locator
relationship
crosswalk_status
control_note

Its central purpose is to prevent the false assumption that the two archive structures are interchangeable.

Crosswalk states include:

VERIFIED — the FBI-native and ZAP/CHS material have been securely joined.

PARTIAL — the general relationship is established, but exact document-by-document or page-by-page mapping remains incomplete.

OPEN — the evidentiary relationship is known conceptually, but the precise native join has not yet been frozen.

NATIVE-ONLY FOR CURRENT SCOPE — later FBI material exists outside the ZAP Sections 1–100 framework presently being cross-consulted.

Representative entries include:

X003 — Stine taxi-cab memorandum.

FBI Vault 1 pp.80–81 ↔ ZAP Section 21.

Status: VERIFIED.

X004 — Stine taxi-cab latent result.

FBI Vault 1 pp.86–88 ↔ corrected ZAP Section 13X.

Status: VERIFIED.

X002 — Stine-letter latent branch.

ZAP Section 11 ↔ exact native FBI page join not yet frozen.

Status: OPEN.

X006 — expanded anatomical exemplar requirement.

ZAP Section 28 ↔ exact FBI Vault 1/2 page join still pending.

Status: OPEN.

X007 — late-November/December 1969 A-10042 screening.

FBI Vault 3 pp.4–19 ↔ multiple reconstructed sections, including Sections 59, 63, 64, and 71.

Status: PARTIAL, because the relationship is not properly represented as a simple one-section/one-page-cluster equivalence.

X011 — FBI Vault 5 upload boundary.

The break between native p.249 and p.250 is recorded as an upload split only. It is not an evidentiary or native FBI section break.

The Native Crosswalk therefore answers:

Where does this reconstructed ZAP/CHS object sit in the surviving FBI-native record, and how certain is that join?

CHS citation form: ZRA9-XWALK X###

Example: ZRA9-XWALK X004

 

B.7 ZRA9_CLAIM_MATRIX.csv — Claim Strength and Publication Control

Function:
The Claim Matrix converts documentary findings into explicit propositions and assigns each proposition a support state.

Its fields are:

claim_id
claim
support_level
supporting_sources
source_state
limitation
publication_safe_formulation

The table is intended to stop a historically supportable observation from expanding, through repetition, into a stronger proposition than the documents warrant.

Examples include:

CL001 — A-10042 as a reusable comparison/search case-set.

Support level: STRONG.

The evidence supports stating that A-10042 was repeatedly used by the FBI as an administrative latent-case set for Zodiac-related comparison and search activity.

It does not establish that every latent within the case-set had the same donor.

CL002 — every latent in A-10042 was deposited by Zodiac.

Support level: NOT SUPPORTED.

Administrative association with the Zodiac investigation does not itself authenticate offender provenance.

CL006 — fingerprint non-identification can coexist with an incomplete comparison universe.

Support level: VERY STRONG.

This is supported recurrently by records requiring tips, sides, lower joints, palms, or later major-case impressions after ordinary fingerprint comparisons had already produced non-identifications.

CL007 — thirty-eight A-10042 fingerprints did not identify Arthur Leigh Allen.

Support level: STRONG.

CL008 — Allen was completely cleared by every relevant Zodiac latent or palm impression.

Support level: NOT SUPPORTED BY AUDITED PACKET.

The safe formulation is that the thirty-eight comparable fingerprints did not identify Allen, while the same report separately describes residual material for which the available exemplar basis was incomplete.

CL010 — A-12397 and A-10042 are the same latent case.

Support level: UNRESOLVED.

CL012 — ALSA3 subnumber count equals the number of unique Zodiac offender fingerprints.

Support level: NOT SUPPORTED.

The Claim Matrix therefore answers:

How strongly can this proposition be stated in publication?

CHS citation form: ZRA9-CLAIM CL###

Example: ZRA9-CLAIM CL007

 

B.8 ZRA9_LOGIC_AUDIT.csv — Source Statement to Permissible Inference

Function:
The Logic Audit is the principal inferential control table.

Its fields are:

logic_id
source_statement_or_fact
permissible_inference
impermissible_escalation
evidentiary_state
source_ref

The file explicitly records the step between what a document says and what the researcher wishes to conclude from it.

For example:

G001
Source fact: thirty-eight previously reported latent fingerprints were not Arthur Leigh Allen's fingerprints.

Permissible inference: those thirty-eight compared latent fingerprints did not identify Allen.

Impermissible escalation: Allen was conclusively excluded as Zodiac by every relevant latent or palm impression.

G002
Source fact: remaining tip- and side-area fingerprints and other impressions were compared with the available fingerprints and no identification was effected.

Permissible inference: the available exemplars did not identify the subject on that residual material.

Impermissible escalation: the residual material definitively excluded the subject.

G003
Source fact: no palm prints were contained in FBI files for the compared individuals.

Permissible inference: a palm comparison could not be completed from FBI-file palms.

Impermissible escalation: palms were compared and did not match.

G004
Source fact: clearly and completely recorded major-case prints were required for conclusive comparison.

Permissible inference: the existing exemplar set was insufficient for conclusive comparison of some questioned material.

Impermissible escalation: an ordinary ten-print card was sufficient to clear all relevant friction-ridge anatomy.

G006
Source fact: Sacramento requested photographs of all latent fingerprints developed during the Zodiac investigation under A-10042.

Permissible inference: A-10042 functioned administratively as a Zodiac-investigation latent case-set.

Impermissible escalation: every print stored or searched under A-10042 was deposited by Zodiac.

G010
Source fact: the FBI could not locate an Identification Division record using supplied identifying information.

Permissible inference: no qualifying Bureau identification record was located using that information.

Impermissible escalation: the individual had no criminal history, or had been fingerprint-cleared in the Zodiac investigation.

The Logic Audit therefore answers:

What does this fact logically permit us to say — and where must the inference stop?

CHS citation form: ZRA9-LOGIC G###

Example: ZRA9-LOGIC G001

 

B.9 ZRA9_ZAP_DESCRIPTION_AUDIT.csv — Secondary Description Control

Function:
The ZAP Description Audit tests the project's reconstructed descriptions of ZAP sections against the underlying documentary findings.

Its fields are:

zap_locator
description_source
project_description_or_label
underlying_document_finding
agreement_status
certainty_risk
action

The current file requires an important qualification.

Its descriptions derive from CHS project state and transcript reconstruction. They are not presently claimed to be verbatim reproductions of original ZAP headings or metadata.

Consequently, a project label such as “Stine cab latent follow-up/result” may be used as an internal descriptive tool where it accurately reflects the underlying document, but it must not be attributed to ZAP contributors as though it were their exact heading unless the original metadata has separately been extracted and frozen.

The table also records places where descriptive shorthand could create evidentiary error.

Section 11 carries a high certainty risk if the Stine-letter branch is merged with the cab result.

Corrected Section 13X must retain cab provenance and must not be described as the Stine-letter latent result.

Section 20 must remain tied to the separate A-12397 Lake Berryessa case unless an administrative join to A-10042 is recovered.

Section 28 must not be shortened to “suspect cleared,” because the underlying document distinguishes ordinary fingerprint non-identification from the need for additional anatomical exemplars.

Sections 59, 63, 64, and 71 preserve the same recurrent distinction.

Section 85 remains specifically marked DEGRADED / RETRIEVAL FLAG and must not be reconstructed beyond what the damaged surviving material supports.

The Description Audit therefore applies to our own archive the same standard demanded of the historical record:

a useful description is not the same thing as the document it describes.

CHS citation form: ZRA9-ZAP [Section locator]

Example: ZRA9-ZAP Section 28

 

B.10 ZRA9_COMPARISON_HISTORY.csv — Suspect Comparison and Exemplar-State History

Function:
The Comparison History records individual and cohort comparison cycles while preserving distinctions among ordinary fingerprint results, palm or extended-anatomy availability, and the availability of usable reference records.

Its fields are:

comp_id
event_date
source_locator
case_id
comparison_subject_scope
request_or_result
ordinary_fingerprint_state
palm_or_extended_anatomy_state
record_availability_state
overall_controlled_disposition

The table exists because the phrase “did not match” can conceal several materially different evidentiary states.

These include:

ordinary fingerprints were available and were non-identical;

ordinary fingerprints were available but palms were not;

ordinary fingerprints were compared while lower-joint or palm-relevant material remained unresolved;

no usable identification record was located;

a request survives but the corresponding result has not been securely recovered;

a result survives but the originating request or subject identity is degraded;

or a comparison was performed but the exemplar set was explicitly insufficient for conclusive examination of all questioned anatomy.

Representative entries include:

P006 — Section 28.

Ordinary comparable prints: non-match.

Extended anatomy: tips, sides, lower joints, and palms required.

Controlled disposition: partial non-match plus explicitly incomplete biometric comparison.

P007 — Section 59.

Available ordinary prints: non-match.

Lower-joint/palm-type impression: unresolved because corresponding standards were absent.

P011 — Section 85.

Status: result-only/degraded.

The exact subject, count, and result language are not reconstructed beyond the surviving material.

P017 — March 29, 1991, redacted subject.

Thirty-seven latent fingerprints did not identify the subject; residual anatomy remained under a different comparison state.

P018 — March 29, 1991, Arthur Leigh Allen.

Thirty-eight latent fingerprints did not identify Allen. Residual tip/side impressions produced no identification with the available fingerprints; FBI-file palms were absent; major-case prints were required for conclusive residual comparison.

The Comparison History therefore answers:

For this person or cohort, what comparison was actually performed, what standards existed, and what is the narrowest accurate disposition?

CHS citation form: ZRA9-COMP P###

Example: ZRA9-COMP P018

 

B.11 ZRA9_UNRESOLVED.md — Open Retrieval and Crosswalk Register

Function:
The unresolved register prevents incomplete questions from disappearing merely because the narrative has reached a provisional conclusion.

It records unresolved work under five categories:

provenance and physical-object identity;

exemplar completeness;

numerical integrity;

FBI-native/ZAP crosswalk completion;

and publication controls.

The file is not an appendix of doubts added after the analysis. It is part of the analytical system itself. A conclusion may only be as strong as the state of the unresolved questions upon which it depends.

Because these open targets materially define the evidentiary limits of ZRA9, they are reproduced in expanded publication form in Appendix C.

CHS citation form: ZRA9-OPEN

 

B.12 ZRA9_A10042_DATA_SPINE.xlsx — Consolidated Analytical Workbook

Function:
The workbook is a convenience interface that consolidates the eight principal analytical CSV datasets into a single spreadsheet environment.

It does not replace the CSV files as canonical tabular objects. Its role is practical: sorting, filtering, comparing, and cross-consulting the source, latent, chronology, crosswalk, claim, logic, description-audit, and comparison-history layers within one working file.

The workbook is therefore best understood as a researcher-facing view of the same underlying data spine.

Where a discrepancy were ever to arise between a workbook view and the separately preserved canonical table, the individually frozen source file and its package hash should control until the discrepancy is reconciled.

CHS citation form: ZRA9-WORKBOOK

 

B.13 ZRA9_PACKAGE_MANIFEST.csv — Package Integrity Record

Function:
The Package Manifest records the preservation state of the ZRA9 data-spine package.

Its fields are:

file_name
bytes
sha256

The manifest contains the state documents, analytical CSV files, unresolved register, and consolidated workbook, together with their byte sizes and SHA-256 hashes.

The purpose of the manifest is not evidentiary interpretation. It is archival integrity.

If a research file is subsequently edited, expanded, corrected, or regenerated, its hash will change. The manifest therefore makes it possible to distinguish the August 16, 2026 DATA SPINE v1 package from later revisions and to verify whether two copies are byte-for-byte identical.

This is especially important for a project in which later crosswalk work may legitimately alter previously open or partial mappings. A future correction should not silently overwrite the historical state of the apparatus. It should produce a new version.

CHS citation form: ZRA9-MANIFEST

 

B.14 How the CHS Apparatus Should Be Cited

A citation to the CHS apparatus should not replace citation to the underlying FBI or preserved ZAP document when that primary source is available.

The preferred hierarchy is:

primary FBI/ZAP object → CHS analytical record → published proposition

For example:

FBI Vault 5, pp.259–263; ZRA9-LATENT L013–L014; ZRA9-COMP P018; ZRA9-LOGIC G001–G004.

Such a citation tells the reader not only where the underlying FBI material can be found, but also which CHS evidence record, comparison record, and inferential control produced the published formulation.

Where a ZAP-preserved document has a stable descriptive filename, that filename may be included as an additional direct locator.

Where the FBI-native/ZAP relationship remains open or partial, the citation should say so rather than presenting an unresolved crosswalk as established.

The CHS architecture therefore permits movement in both directions:

article claim → CHS analytical record → preserved source object

and

preserved source object → CHS record → publication claim

This bidirectional traceability is the principal purpose of the data spine.

 

B.15 What the Data Spine Is — and Is Not

The ZRA9 data spine is not intended to manufacture certainty through database structure.

A row identifier does not make an interpretation true.

A crosswalk does not create a historical relationship that the documents themselves do not establish.

A case number does not identify a donor.

A search-record count does not become a unique-print count merely because it can be enumerated.

And a later project description does not become FBI language merely because it accurately summarizes an FBI document.

The function of the system is instead to make uncertainty explicit, stable, searchable, and reviewable.

Its fundamental rule is therefore the same rule that governs ZRA9 as a whole:

the archive must preserve not only what the evidence establishes, but the point at which the evidence stops.

 

 

APPENDIX C — OPEN QUESTIONS AND RETRIEVAL REGISTER

C.1 Purpose

The reconstruction presented in ZRA9 is sufficiently developed to support a number of strong conclusions, but it is not complete.

The remaining gaps are not treated as peripheral imperfections to be concealed beneath a finished narrative. They are part of the evidentiary result.

A forensic archive should be capable of saying not only what has been established but also which questions remain unanswered, what documentary object would be required to answer them, and which current conclusions depend upon those unresolved relationships.

The following register therefore preserves the open research targets controlling the next phase of ZRA9 and the wider CHS reconstruction.

 

C.2 Priority One — Provenance and Physical-Object Identity

C.2.1 Recover the Final FBI Disposition of the Stine-Letter Latents

The current archive establishes a distinct October 17, 1969 documentary branch in which SFPD reported latent impressions recovered from the October 13 Stine letter and indicated that photographs would be supplied for FBI searching.

What remains missing is an explicit final FBI document securely joining that submission to a later examination result.

Until such a document is recovered, the Stine-letter latent branch must remain separate from the Stine taxi-cab packet.

The October 19 cab memorandum and October 29 cab result cannot be used as substitutes merely because both bodies of evidence arise from the Paul Stine investigation.

Retrieval target: the FBI receipt, work sheet, search record, report, return correspondence, or other document explicitly disposing of the October 13 letter-latent photographs.

Publication control: do not state that the known Stine-cab latent result is also the documented result of the Stine letter.

 

C.2.2 Crosswalk the 1991 Q2, Q3, and Prior Q16 References to Physical Evidence Objects

The 1991 reactivation refers to Q2, Q3, and a prior Q16. At least one prior Q16 impression is described as originating from a lower-joint area of a finger or possibly from a palm.

The surviving archive does not yet establish the complete physical-object identity behind those designations.

This matters because Q and Qc numbers are case-context identifiers. The appearance of “Q16” in two separate documentary contexts does not, by numerical similarity alone, prove that the same physical object is being referenced.

Retrieval target: original submission schedules, laboratory worksheets, transmittals, evidence inventories, or later cross-reference documents connecting each 1991 Q designation to its carrier object.

Publication control: do not equate the 1991 prior Q16 with the August 1969 Concerned Citizen Q16 envelope unless the object-level crosswalk is documented.

 

C.2.3 Map A-10042 Entries to Their Underlying Source Objects

A-10042 can now be shown to have operated as a long-lived administrative comparison and search case-set.

What has not yet been reconstructed impression by impression is the source-object architecture of the entire case.

We do not yet possess a definitive table establishing which A-10042 latent entries derive from the Stine cab, correspondence, other questioned documents, other physical evidence, or potentially other investigative submissions.

Without this object-level map, administrative membership in A-10042 cannot be converted into donor attribution.

Retrieval target: original A-10042 evidence schedules, latent-photo indexes, Identification Division worksheets, consolidation records, submission correspondence, or any surviving object-to-latent inventory.

Publication control: describe A-10042 as an administrative Zodiac-investigation latent case-set, not as a homogeneous set of impressions conclusively established to have been deposited by one offender.

 

C.2.4 Determine the Administrative Relationship Between A-10042 and A-12397

Lake Berryessa records preserve the separate latent-case designation A-12397.

Elsewhere, A-10042 emerges as the long-lived latent comparison and search case-set associated with the wider Zodiac investigation.

The current archive has not recovered a primary administrative document establishing whether A-12397 was later incorporated into, cross-referenced with, consolidated into, or merely compared against A-10042.

Similarity of investigative purpose is insufficient to answer that question.

Retrieval target: FBI case-opening records, consolidation memoranda, latent-case indexes, evidence-transfer forms, or Identification Division administrative correspondence joining or distinguishing the two case numbers.

Publication control: preserve A-12397 and A-10042 as separate printed FBI identifiers unless and until their administrative relationship is proven.

 

C.3 Priority Two — Exemplar Completeness and Elimination History

C.3.1 Recover Later Major-Case Prints for Arthur Leigh Allen and Other 1991 Subjects

The March 29, 1991 report states that clearly and completely recorded major-case prints were required for conclusive comparison of residual tip-, side-, lower-joint-, palm-relevant, or other extended-anatomy material.

The present archive does not yet establish whether such major-case cards were subsequently obtained for Arthur Leigh Allen or the other subjects involved in the 1991 comparison cycle.

Nor does it establish whether a later conclusive residual comparison was performed after such exemplars, if obtained, became available.

Retrieval target: subsequent Identification Division correspondence, major-case fingerprint cards, palm cards, follow-up laboratory reports, state or local submissions, and post-March 1991 comparison worksheets.

Publication control: preserve the March 29 result exactly as documented: thirty-eight comparable fingerprints did not identify Allen; the residual comparison universe described in the same report had a different and incomplete exemplar state.

 

C.3.2 Recover Palm and Elimination Exemplars for Known Potential Donors

A questioned latent impression acquires substantially greater evidentiary meaning when known alternative donors can be excluded.

The present archive does not yet contain a complete elimination-print universe for the relevant physical environments and handled evidence.

Potential known donors may include, depending upon the object and handling chain:

victims;

police officers and other first responders;

postal personnel;

newspaper personnel;

laboratory personnel;

cab-company personnel;

mechanics or cleaners;

law-enforcement evidence handlers;

and other documented persons with legitimate access to the carrier object or crime-scene surface.

The existence of an unidentified impression is not equivalent to the identification of an offender if realistic alternative donors have not been exhaustively eliminated.

Retrieval target: elimination fingerprint cards, palm cards, evidence-handling logs, local police latent files, laboratory elimination worksheets, and related correspondence.

Publication control: unidentified remains unidentified unless donor attribution is independently established.

 

C.3.3 Determine Whether Residual A-10042 Impressions Can Be Eliminated to Known Non-Offender Donors

The recovery of a latent impression and the failure to identify it to a suspect are only part of the donor question.

A separate question is whether some presently residual impressions could be attributed or eliminated to known non-offender contributors.

The Stine cab demonstrates the importance of this process because at least one submitted latent fingerprint was identified as Paul Stine's.

The same evidentiary discipline must be applied to other questioned impressions.

Retrieval target: later identification results, local latent comparison files, elimination records, and donor-reference collections.

Publication control: absence of a suspect identification does not by itself authenticate the impression as offender-origin.

 

C.4 Priority Three — Numerical Integrity

C.4.1 Resolve the Relationship Between the Thirty-Seven- and Thirty-Eight-Latent Groups

The March 29, 1991 report records thirty-seven previously reported latent fingerprints as not belonging to one redacted subject and thirty-eight latent fingerprints as not belonging to Arthur Leigh Allen.

The numerical difference is real.

Its explanation is not yet established.

The groups may represent the same underlying case-set with different comparability for the two subjects; overlapping but non-identical subsets; different available reference records; an added or excluded latent; or another administrative or examination distinction.

The present archive does not justify choosing among those possibilities.

Retrieval target: complete examiner worksheets, latent-number schedules, subject-specific comparison sheets, or a latent-by-latent 1991 inventory.

Publication control: do not describe the thirty-seven and thirty-eight groups as identical, wholly different, or directly additive until the relationship is documented.

 

C.4.2 Build a Record-Level ALSA3 Subnumber Crosswalk

The February–March 1991 material contains a large corpus of A-10042 ALSA3 subnumbers and automated-search records.

These records demonstrate substantial computerized processing.

They do not yet establish a unique-latent count.

A subnumber may represent an encoded latent, an individual automated search, a repeated search, a duplicate, a consolidation, an administrative transaction, or another information-system operation.

Retrieval target: record-level ALSA3 documentation linking each subnumber to the questioned latent, image, search event, disposition, and any duplicate or consolidation relationship.

Publication control: do not convert the number of ALSA3 subnumbers into the number of unique Zodiac fingerprints.

 

C.4.3 Resolve Consolidation and Non-Deletion Records Within the ALSA3 Corpus

Some ALSA3 records appear to reflect consolidation or administrative handling rather than independent latent evidence.

Those relationships must be reconstructed before numerical claims about the case-set can be considered secure.

Retrieval target: ALSA3 coding instructions, consolidation records, search-history documentation, deletion/non-deletion records, and latent-image references.

Publication control: information-system records must remain distinguishable from physical impressions.

 

C.5 Priority Four — FBI Vault / ZAP / CHS Crosswalk Completion

C.5.1 Freeze Exact Native FBI Pages for Key ZAP Sections

The content of several reconstructed ZAP sections is sufficiently understood for controlled analytical use, but their exact native FBI Vault page joins have not yet all been frozen.

Priority sections are:

Section 28
Section 59
Section 63
Section 64
Section 71
Section 85
Section 86
Section 87
Section 92
Section 93

These sections are especially important because they preserve the recurring A-10042 comparison architecture: ordinary-print non-identification, record availability, lower-joint/palm limitations, request/result chains, and degraded or incomplete packets.

Retrieval task: locate each underlying document in the native FBI Vault sequence and record the exact one-to-one, one-to-many, or many-to-one page relationship.

Publication control: where the native join remains unresolved, cite the ZAP/CHS locator honestly rather than manufacturing a precise FBI page citation.

 

C.5.2 Preserve One-to-Many and Many-to-One Relationships

Crosswalking is not an exercise in making two archives appear identical.

If one ZAP section contains material drawn from several FBI-native pages or document clusters, that relationship must be preserved.

If several reconstructed sections divide material belonging to one FBI-native administrative packet, that relationship must also be preserved.

The same rule applies when a document crosses a native storage boundary or when a later upload break has no evidentiary significance.

Publication control: never force “ZAP Section N = FBI pages X–Y” merely for convenience where the documentary relationship is more complex.

The crosswalk is intended to describe the archive, not simplify it beyond recognition.

 

C.5.3 Extract and Freeze Original ZAP Metadata Separately

The current ZRA9_ZAP_DESCRIPTION_AUDIT.csv uses CHS/WERK-FBI project descriptions derived from the reconstructed state of the archive.

Those descriptions are analytical labels.

They are not currently claimed to reproduce original ZAP section headings, captions, notes, or contributor metadata verbatim.

A separate metadata pass is therefore required.

Retrieval target: original ZAP headings, filenames, captions, section descriptions, dates, contributor notes, and other available metadata in their original form.

Publication control: clearly distinguish:

original ZAP metadata

from

CHS/WERK-FBI descriptive labels

from

FBI source language

from

project evidentiary assessment.

This is the same provenance discipline applied to our own archival work that ZRA9 applies to the FBI record.

 

C.6 Priority Five — Permanent Publication Controls

C.6.1 Preserve the Difference Between Certain Result and Uncertain Evidence Universe

Every published table, article section, suspect comparison, and future ZRA analysis should continue to distinguish:

the certainty of the comparison result;

the provenance of the questioned impression;

the anatomical area represented;

the completeness of the reference exemplar;

the elimination status of known alternative donors;

and the completeness of the evidence set being generalized from.

A definite fingerprint non-identification must not be weakened merely because other evidence remains unresolved.

Conversely, a definite non-identification of a defined subset must not be expanded into a universal exclusion when the same packet preserves unresolved anatomy or absent exemplar classes.

This is not a temporary caution applicable only to Arthur Leigh Allen.

It is a permanent rule of the CHS archive and must be applied equally to every subject.

 

C.6.2 Keep Evidentiary Lanes Separate Unless a Primary Document Joins Them

The FBI Zodiac archive contains multiple forensic and investigative disciplines.

These include:

latent fingerprints and palms;

questioned-document and handwriting examination;

cryptanalysis;

voice evidence;

Identification Division record searches;

administrative memoranda;

authentication inquiries;

and other investigative intelligence.

The existence of information in one lane does not automatically modify the evidentiary status of another.

A handwriting assessment is not a fingerprint result.

A failure to locate an FBI Identification record is not a latent-fingerprint non-match.

A cryptanalytic conclusion is not an authentication of a physical latent.

An administrative decision to continue or close an investigative lead is not itself a forensic examination.

Publication control: evidentiary lanes may be analytically compared, but they should only be formally joined when a primary document establishes the relationship.

 

C.7 Governing Open Question

The unresolved register ultimately returns to the same problem that governs ZRA9:

What can be said with certainty about a comparison when the donor identity, anatomical coverage, or exemplar completeness of the evidence universe remains uncertain?

The answer is not that certainty is impossible.

Nor is it that every negative comparison is weak.

The correct answer is narrower.

A comparison may be definitive with respect to the impressions and exemplars actually compared while remaining silent about evidence that was not comparable, exemplars that did not exist, anatomical areas that were not represented, donors who were never eliminated, or questioned impressions whose offender provenance was never established.

The task of the archive is to preserve those boundaries rather than erase them.

 

C.8 Status of the Register

These questions should remain designated OPEN until documentary evidence permits an individual item to be resolved.

Resolution should not delete the historical question from the archive.

Instead, a resolved item should be transferred into the appropriate source, latent, chronology, crosswalk, comparison, claim, or logic record, with the former unresolved state preserved in version history.

The unresolved register is therefore not evidence that the reconstruction has failed.

It is evidence that the reconstruction has identified exactly where the surviving record ceases to support a stronger conclusion.

That boundary is itself a finding.