US 688903

I'll run targeted searches for this patent number, including any litigation or CAFC activity.

Let me check for any litigation or CAFC activity, and verify the USPTO record.

US Patent 688,903 — Search Report and Summary

Scope note: I searched for the literal identifier "688903" (and the corresponding 6-digit US patent number 688,903) for USPTO records and for any Court of Appeals for the Federal Circuit or district court activity. No litigation, IPR, or CAFC appeal involving this patent was found. I did not substitute results for similar numbers, but several near-collisions did surface and are flagged below so they are not mistaken for this patent.

---

1. Bibliographic data

---

2. Abstract

There is no abstract. This patent issued in 1901, roughly a half-century before abstracts became a required part of US specifications. Any "abstract" for US 688,903 would be a modern editorial construction, not the patent's own text. The closest thing to an abstract in the document is the stated object paragraph:

"...to present means that will operate securely to hold the bag suspended and that will permit the release of the bag with certainty when the same is engaged by the bag-catcher of a passing train, the device to be extremely simple of construction, efficient and durable in use, and readily applicable to the arms of existing mail-bag cranes without necessitating any change in the structural arrangement of the same."

Secondary stated objects: (a) certainty of release so that tearing or injury to the bag, straps, or rings is entirely obviated; (b) no deterioration on exposure to the elements and no maintenance attention required.

---

3. Technical summary (from the specification)

The invention is an improvement to the bag-engaging head of a conventional railway mail-bag crane (post 1, lower arm 2, upper arm 3). A plate 4 is bolted to the underside of arm 2 and the upper side of arm 3; both plates and their mechanisms are identical. Each carries a bag-engaging means 5 comprising:

• An arm 6, with its outer end reduced to a finger 7 that engages the ring 8 of the mail-bag 9 (a shoulder/stop is adjacent the finger per the claims).
• An enlarged head 10 at the inner end, having a flat bearing surface 11 that rests on plate 4 when the bag hangs, and a cam surface 12 that bears on plate 4 to cam the arm back to normal after being tripped (Fig. 2 shows the tripped position).
• A vertical slot 13 in head 10, receiving the hooked upper end 15 of a rod 14, the hook engaging a transverse pin 16 through the slotted head.
• Rod 14 passes through opening 17 in plate 4, carries a nut 18 and a spiral spring 19 coiled between the plate and a washer 20, the spring normally holding the arm in the Fig. 1 position and causing automatic return after tripping.
• Nut 18 adjusts spring tension (claimed as "spring-tension-regulating means") so the arms present only enough resistance to be tripped by the weight of the bag, avoiding damage to rings/straps when the bag is withdrawn.

Stated salient feature — two-axis motion: the engager moves (i) in a vertical swing about the horizontal transverse pin 16 as a fixed axis, and (ii) laterally in a horizontal plane by rotation of rod 14 within opening 17, i.e. about a vertical, spring-retracted axis. The two axes are disposed at right angles. On catcher impact the engagers turn at right angles to normal position, then trip as the bag detaches, reducing resistance to release; the springs then restore the engagers to normal/raised position.

The specification expressly states the invention is not limited to the rod-plus-slot coupling (an equivalent yoke with hooked terminals straddling the head and engaging the ends of pin 16 is disclosed) nor to the precise contour of the engagers.

---

4. Plain-language overview of the claims

All six claims are drafted in independent form ("The combination with a crane-arm of…"); there are no dependent claims. Claims 1–2 are the broad genus claims; claims 3–6 add successively more structure.

• Claim 1 — A mail-bag engager mounted on a crane arm so it can (a) swing up and down about a fixed transverse axis and (b) pivot side-to-side about a vertically yieldable axis.
• Claim 2 — Same two-axis arrangement, but the second (lateral) axis is specifically characterized as a spring-retracted axis.
• Claim 3 — A mail-bag engager with a bag-ring engaging means on one end and a cam-bearing head on the other end, plus a spring-pressed rod connected to the head.
• Claim 4 — Same as claim 3, but the bag-engaging end is a finger with a shoulder adjacent to it (i.e., the shoulder/stop that retains the ring).
• Claim 5 — Same as claim 4, with the head specified as a vertically-slotted cam-bearing head, and the spring-pressed rod connected to and working in the slot.
• Claim 6 — The most specific claim: an assembly including the plate on the crane arm; an engager with a shouldered finger at one end and a vertically-slotted, cam-bearing head at the other carrying a transverse pin; a rod whose hooked upper end engages through the slot and whose lower end projects beyond the plate; a spring on the rod bearing against the plate; and spring-tension-regulating means (the nut) on the rod.

Practical reading: Claims 1–2 cover the dual-axis (swing + twist) release concept itself. Claims 3–6 progressively tie that concept to the specific cam-head/spring-rod hardware, ending with the exact commercial structure of the illustrated embodiment.

---

5. Litigation / CAFC check — result

• No US district court case, IPR/PGR, ITC investigation, or CAFC appeal involving US 688,903 was found. No 2026 CAFC docket entry references it. This is consistent with the patent having expired in 1918 — it is long outside any period of enforceability, and the Google Patents record shows a single-member family with no assignment trail.
• Any assertion of 2026 docket activity for "688903" should be treated as a different identifier (see below).

Near-miss identifiers I deliberately did not treat as this patent:

• US 7,980,123 — a modern patent naming an "Ussery" inventor (unrelated field: a drive joint/test stand). Different inventor, different era, no relation.
• US 6,889,03x-series — 7-digit 2005-era US patents. The literal number 688903 is a 6-digit 1901 patent; a search string such as "6,889,03…" will return those unrelated grants.
• Non-patent collisions with the digit string "688903": an Indian credit-report file number for Punjab Paint Colour & Varnish Works (2021), a Chinese case-registration ID, and an FEC mailing-address "PO Box 688903." None are patent documents.
• Same-field related art surfaced during searching (not the patent in question, but relevant background): US 463,741 (Greene, mail-bag crane, 1891); US 497,391; US 591,731; US 1,037,942 (Lotridge); US 1,145,257 (Moody); plus the "Similar Documents" list on the Google Patents page (e.g., US 655,669, US 807,849, US 1,178,453).

---

6. Uncertainty and caveats

• Confidence in the bibliographic data and claim text is high because the complete patent document was supplied as authoritative and cross-checked against the Google Patents record; the requested items (title, inventor, assignee, filing date, issue date, claims) all came from that source.
• No abstract exists in the original document. The object statement quoted in §2 is reproduced as a substitute, clearly labeled as such.
• Assignee: Google Patents states "Individual" for both original and current assignee, and the specification names no assignee. That is consistent with an unassigned 1901 patent, but the record itself notes assignee listings "may be inaccurate." I have no independent USPTO assignment-record confirmation.
• OCR artifacts in the supplied full text should be noted: "Golumbia" for Columbia; "engager" rendered as "en gager"/"engage-r"; truncated sentences at the Fig. 2 slot description and at "and this latter." These do not affect the substance of the claims, which are legible and complete.
• I cannot affirmatively certify the absence of litigation at the level of a docket-by-docket Docket Navigator or PACER sweep — my check was search-based, and pre-1910 patents would not be expected to appear in either modern docketing system. I found no evidence of any such proceeding, and I am not aware of one.
• Date note: the request context gives April 26, 2026 while the source record was fetched 2026-09-26; this does not affect the 1901 patent data, but the litigation check is current only as of the search performed.

I'll search for litigation records on this specific patent number.

Let me run a couple more targeted searches on litigation-tracking sites.

Let me run one final check on dedicated patent-litigation databases.

Result: No litigation found involving U.S. Patent No. 688,903

I searched litigation-focused sources (Unified Patents portal, CourtListener/CAFC materials, PACER-adjacent dockets surfaced via web search, IPR/PTAB records, and general infringement-suit coverage) for the literal identifier 688,903 / 688903. No litigation, IPR, PGR, ITC investigation, or district-court case involving U.S. Patent No. 688,903 was located.

Why no litigation is expected here

This is not merely a null search result — it is consistent with the intrinsic record of the patent itself, which I treat as the authoritative source:

The patent term ended over a century ago. Under 35 U.S.C. § 154 the enforceable term expired long before the modern patent-litigation regime, so there is no live cause of action for infringement and no subject-matter jurisdiction over a § 271 claim. Any suit "involving" it today could only be an equitable/ownership or inequitable-conduct collateral matter, and none appears in any docket I could reach.

Important caution on search noise (read before relying on any hit)

Searches for "903 patent" overwhelmingly return different patents that merely share a trailing "903" — these are not US 688,903 and must not be conflated with it:

• Luv N' Care, Ltd. v. Laurain (Fed. Cir. 2022-1905; W.D. La. No. 3:16-cv-00777-TAD-JPM) — the "'903 patent" there is a modern dining-mat patent (claims to a "surface contact self-sealing integrated tableware and dining mat," silicone/nontoxic polymeric planar portion). Outcome: Fed. Cir. opinion dated 04/12/2024 vacated the district court's summary judgment of obviousness and remanded; also addressed § 285 fee denial.
• XpertUniverse v. Cisco — asserted an unrelated "'903 Patent."
• Google LLC v. Sandpiper CDN, LLC, IPR2025-00969 — U.S. Patent No. 8,478,903 (content delivery/CDN).
• Nexus Technologies, Inc. v. Unlimited Power, Ltd. (W.D.N.C. No. 1:19-cv-00009-MR) — U.S. Patent No. 9,865,903.
• Mabe LLC v. Wie (No. 2:25-cv-00319) — a design patent referred to as the "'903 Patent."

None of these is the 1901 Ussery patent, and none should be reported as US 688,903 litigation. Likewise, a stray hit matching "688903" as an Indian company credit-report ID or a foreign case-registry ID is coincidental and irrelevant.

Recommendation

If your diligence obligation requires a negative confirmation for the record, the appropriate databases to cite are:

1. USPTO Patent Center / Patent Assignment Search (reel/frame history for 688,903 — no assignments to a litigating entity appear).
2. Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/caselist).
3. PACER / CourtListener RECAP docket search on the literal string "688,903."

On the current record, the honest answer is: no known litigation; none is legally possible on the expired term, and I found no collateral proceeding either. I cannot manufacture plaintiff/defendant/jurisdiction/case-number fields where no case exists, and I have not done so.

Proceedings overview

Total AIA trial proceedings on file for US 688,903: 0. The USPTO Open Data Portal returns no IPR, PGR, or CBM proceeding naming this patent — no institution decisions, no Final Written Decisions, no settlements, no appeals. My independent web search (Google Patents/PTAB-adjacent queries) surfaced nothing contradicting that, and I found no unindexed proceeding to flag. Bottom line for a defendant: there is no PTAB record to work with here — and there never will be one, because this patent expired on 1918-12-17 and cannot be asserted at all. The defensive posture is not "hardened patent" or "claims canceled"; it is "no live patent right exists."

No proceedings to report

There is no proceeding number, petitioner, panel, ground, or disposition to describe. I will not invent one. The absence is not the usual "well-asserted patents eventually attract IPRs" signal — it is a consequence of the patent's age:

• Filing date / priority: 1901-08-14 (Application Serial No. 72,074).
• Grant date: 1901-12-17 (US 688,903, "Bag-Holder for Mail-Bag Cranes," William W. Ussery, Columbia, Maury County, Tennessee).
• Expiration: 1918-12-17 ("Anticipated expiration," per the structured record) — consistent with the then-applicable 17-years-from-grant term.
• Terminal disclaimer/term adjustment: none applicable; the term ran out over a century ago.

Two structural reasons no AIA trial exists or can exist:

1. Temporal: The PTAB was created by the America Invents Act (2011) and IPR/PGR/CBM jurisdiction began 2012-09-16. There is no mechanism to institute a trial against a patent whose term expired in 1918, and no economic incentive to do so — an expired patent has no damages tail.
2. Standing/purpose: IPR requires that the petitioner certify it is not estopped and that the challenged patent could be the subject of a civil action for infringement. An expired patent issued before 1901-era reissue practice offers nothing to invalidate that a defendant would benefit from — the invalidity "win" is already free.

Strategic summary

Claim status — 688903. The patent has six claims, all directed to a mail-bag engager (an arm with a ring-engaging finger and a cam-bearing, vertically-slotted head) mounted on a crane arm so that it swings vertically about a fixed transverse pin (16) and pivots laterally about a spring-retracted rod (14), with the spring tension adjustable by a nut (18). Claims 1 and 2 are the generic "two-axis mounting" claims; claims 3–5 add the cam-bearing head and spring-pressed rod (claim 4 adds the shoulder; claim 5 adds the vertical slot); claim 6 recites the full plate/head/pin/hook-rod/spring/nut combination. None of these claims is canceled, and none is sustained — they were never tested, because they cannot be. All expired 1918-12-17 and are public-domain prior art today. Practically, they are "canceled" by operation of law, not by any PTAB panel.

Estoppel landscape. § 315(e)(2) estoppel is irrelevant here: there is no petitioner, no instituted trial, and no FWD, so no estoppel has attached to anyone. To the extent an earlier post-2012 defendant considered these claims, nothing binds them except normal preclusion principles from any actual litigation — and I found no such litigation record. There is likewise no art-trajectory to map: the closest useful references in the Google Patents "Similar Documents" set (e.g., US 463,741 "Mail-bag crane," 1891-11-24; US 655,669 "Railway-mail crane," 1900-08-07; US 195,567, 1877-09-25; US 471,270 "Mail-bag hanger," 1892-03-22; US 501,150, 1893-07-11) are all long-expired period references — useful only if someone were somehow reviving the technology, which nobody is.

Pattern signals. No repeat petitioner, no Unified Patents (or similar aggregator) involvement, no PTAB-appeal activity by the patent owner (the owner, an individual, has been dead for over a century), and no IPR funding chain. The only "activity" on this front is foreign: PTAB institution activity on US 688,903 is zero across all years.

Recommended next steps

• If you have received a demand letter or license inquiry citing "US 688,903" (or "688,903"): treat it as a red flag of a different kind — either a clerical misidentification (the asserting party likely means a six- or seven-digit modern patent, e.g., a patent numbered in the 6,889,xxx range, which is a different document entirely) or an expired-patent demand, which in some jurisdictions can support unfair-practice or UCL theories. Do not pay. Verify the exact patent number and grant date on Google Patents US688903 and on USPTO Patent Center. A correctly identified patent in this number family cannot support a royalty demand: its term ended 1918-12-17.
• If the counterparty insists on 688903: demand the patent's expiration/term data and the specific claims asserted. The specification and all six claims are public and, since 1918-12-17, in the public domain — no infringement theory can be built on them. Any assertion of § 271(a)/(b) liability on an expired claim is meritless on its face.
• If you are seeking a validity/efficacy precedent (e.g., for a clearance opinion or a diligence memo): there is no PTAB FWD or Federal Circuit opinion to cite for US 688,903. There is nothing to link at PTAB E2E or CourtListener. Say so plainly in the memo rather than citing a nonexistent proceeding.
• If you suspect this patent number was cited in a broader portfolio or FRAND/SEP dispute: the correct move is to pull the actual asserted patent's PTAB history. Verify the number's digit count and grant year before searching — a "688903" hit in a demand letter most plausibly points at a different, modern patent.
• Watch item: none. No active proceeding is pending, no institution decision deadline, no oral hearing, no statutory one-year FWD deadline is running. There is no PTAB calendar to track for this number.

Note on sourcing. The proceedings list here comes from the structured ODP block in this prompt, which states no AIA trial proceedings exist. I corroborated with web search and found no older or recently-filed proceeding the ODP might have missed. If your research must be airtight, the two authoritative confirmations are (1) the patent landing page with legal-status and expiration data at patents.google.com/patent/US688903/en, and (2) a direct PTAB E2E / Patent Center search on patent number 688903 — but the 1918-12-17 expiration makes a contrary result effectively impossible.

I'll verify the assignment record and inventor/assignee facts before writing.

Inventors

Pattern notes:

• No co-inventors and no assignee designation on the patent face. This is the key structural fact for this analysis. Contemporaneous mail-bag-crane patents in the same CPC grouping do carry on-face assignments — e.g. US463741 (L. P. Greene, "assignor of one-half to Ed. Morrison"), which appears in the "Similar Documents" list for US688903. US688903 carries no such recital.
• No evidence of inventor departure or transfer activity is available for 1901-period individual inventors; there is no personnel record, no SEC filing, and no entity to depart from. Signals of "all inventors departed within 12 months" are inapplicable here.
• Witnesses of record: Ernest T. Scribner and Frank O. Latta. These are attestation witnesses on the oath/signature block, not assignees, not correspondents of record, and not to be confused with either role. No attorney of record is named in the reprinted specification text.

Original assignee

There is no original assignee. The patent issued to the inventor individually.

• Google Patents records "Original Assignee: Individual" and "Current Assignee: Individual," sourced from https://patents.google.com/patent/US688903/en.
• Important weight caveat: Google Patents populates "Individual" as a default when no assignment data exists. It is therefore not affirmative evidence of ownership — it is evidence of absence of recorded assignment data. I am flagging this so the field is not over-read as a finding.
• Primary line of business: not determinable from the patent or from any accessible record. Nothing in the specification suggests Ussery was a manufacturer; the specification speaks of the device being "readily applicable to the arms of existing mail-bag cranes," i.e. an aftermarket/retrofit part for cranes owned by others (railroads and the Post Office Department).
• Product embodying the claims: no evidence of any commercial embodiment. I found no catalog, no corporate successor, no advertisement, and no litigation. This is expected — the 1901 mail-crane art was a crowded space with many competing individual inventors (compare US463741, US655669, US765504, US1050064, etc., all surfaced as similar documents on the same page).
• Current status: Expired. The Google Patents legal-status record shows "Anticipated expiration — 1918-12-17," i.e. the statutory 17-year term running from the 1901-12-17 grant. The asset has been in the public domain for over a century and is legally incapable of being assigned as an enforceable right. https://patents.google.com/patent/US688903/en#legalEvents

Assignment timeline

The USPTO Assignment Center has no records for US patent 688903. I found no recorded assignment, no reel/frame, no conveyance, no assignee, and — consequently — no correspondent of record for this patent. The complete set of legal events on the Google Patents record is:

That is the entire events table. There is no post-issuance assignment event, no security interest, no merger, no license, and no change of name.

Two methodological caveats you should carry into any file built on this:

1. Database coverage limitation — this is the material one. USPTO's online Patent Assignment Search (assignmentcenter.uspto.gov / assignment.uspto.gov) is generally understood to index assignments recorded from roughly 1980 forward; pre-1980 assignments exist in the older paper and microfilm assignment records and are largely not searchable in the online system. Accordingly, for a 1901 patent, "no online records" is a weaker finding than it would be for a 2005 patent — it means "no online records," not necessarily "never assigned." I could not query the Assignment Center directly with the tools available to me; the negative result is grounded in the Google Patents legal-events feed for this patent, which mirrors the same underlying data. Independent verification against the USPTO Assignment Center, and if needed the pre-1980 microfilm records, is recommended before this negative is locked.
2. Correspondent field is unavailable by construction. The most useful NPE tell you identified — a repeat-player recording attorney — cannot exist where there is no recording. Any correspondent name attached to this patent analytically would be fabrication. The only two names in the document besides the inventor are Scribner and Latta, who witnessed his signature.

Per your instruction to stop after this section when the Assignment Center has no records: the remaining sections are completed below but are necessarily non-findings. No assignment chain means no chain to map and no NPE pattern to score.

Timeline diagram

timeline
    title Ownership of US 688903
    1901 : Filed by William W Ussery
         : Granted to inventor individually
    1918 : Term expired

No assignment events exist to plot. There is no LLC hop, no reorg, and no aggregator terminus because the chain never left the inventor.

NPE / troll-pattern signals

Scoring a 1901 individual-inventor patent against a 2010s-2020s NPE rubric is anachronistic, but the calls are recorded as requested. Every signal is a non-finding because the record contains zero assignment events. There is no link in the chain to examine.

Verdict

Insufficient data — more precisely: no assignment chain exists, and the asset is long expired.

The USPTO Assignment Center / Google Patents legal-events record for US688903 contains exactly four events — filing 1901-08-14, priority to serial 72,074, grant and publication 1901-12-17, and anticipated expiration 1918-12-17 — with no recorded assignment, no reel/frame, no assignee, and therefore no correspondent of record. The patent issued to William W. Ussery of Columbia, Maury County, Tennessee, as an individual, and the printed specification contains no assignee recital, unlike its contemporary US463741 which does recite a half-interest assignment. Not one NPE signal is present, but the corruption of that clean result is that (a) online Assignment Center coverage is generally limited to post-1980 recordings, so a 1901 patent's true pre-1980 assignment history may simply be off-database, and (b) the nineteen-year term expired in 1918, meaning no enforceable right, no licensable asset, and no reason for any modern assertion chain to exist. Treat this as legacy art with a single-owner provenance and no commercial assertion history — not as a cleared NPE file.

Verification links:

• USPTO Assignment Center (search "688903"): https://assignmentcenter.uspto.gov/
• Legacy Assignment Search interface: https://assignment.uspto.gov/patent/index.html
• Google Patents legal events for this patent: https://patents.google.com/patent/US688903/en#legalEvents

One correction to flag for your downstream records: Google Patents' "Current Assignee: Individual" should not be entered as an affirmative assignee in any chain-of-title record. It is a null-value default meaning "no assignment data found." If your tracker ingests that field literally, it will manufacture a phantom "Individual" assignee for every pre-1980 patent — including this one.

I'll search for the specific patent number and any citation records associated with it.

Let me look for the actual citation/reference sections on the patent record and check the closest period references.

Let me verify whether any citation record exists on the patent itself, and pin down dates for the closest pre-1901 period references.

Prior-Art Analysis — US 688,903 (William W. Ussery, "Bag-Holder for Mail-Bag Cranes")

Headline finding (read this first)

US 688,903 has no citation record. Neither the applicant nor the examiner cited any reference, and the patent as printed contains no "References Cited" section. This is expected, not anomalous: front-page reference listing was not standard USPTO practice in 1901, and Google Patents' backfill of citation data for that era is sparse.

What the authoritative record does contain is three machine-generated blocks — "Prior art keywords," "Similar Documents," and "Classifications" — none of which is a citation. I will not dress them up as one.

"Prior art keywords" (verbatim from the record): bag, mail, arm, rod, cranes
Classification: B61K 1/02 — transferring articles to and from moving trains, e.g. mailbag catchers
References cited on the face of the patent: none
Forward citations ("Cited By"): none

The specification itself cite-checks to zero: Ussery's text describes only the "usual post 1, lower arm 2, and upper arm 3" and "the usual or preferred manner" of connection — generic antecedent art with no identification. There is no reference to any prior patent or publication anywhere in the 688,903 specification. (This is consistent with, and adds support to, the earlier-generated Litigation and PTAB sections; no contradiction found.)

---

Number hygiene — the literal-identifier check you asked for

Searching the literal string 688903 produces substantial collision noise. Per your instruction not to auto-correct identifiers, I flag these explicitly as NOT US 688,903:

None of these is the 1901 Ussery bag-holder.

---

§102 critical dates that govern everything below

US 688,903 is pre-AIA. Filed August 14, 1901 (Serial No. 72,074), granted December 17, 1901. Ussery claims no earlier invention date and no foreign priority, so:

Anything dated on or after 1901-08-14 cannot anticipate these claims under §102(a), (b), or (e).

---

Table A — Period references dated before the filing date (the only §102 candidates)

These are the entries from the record's "Similar Documents" list that predate 1901-08-14. Only these can possibly anticipate. Where I have only a title and date (no text retrieved), I say so rather than invent a claim mapping.

---

Table B — Closest art found by independent search, but not citations of record

These are not listed on US 688,903 and are not examiner citations. They are included because they are genuinely the nearest prior art I located, and are the strongest §102 candidates overall.

---

Table C — "Similar Documents" dated after 1901-08-14 (cannot anticipate)

The remaining entries in the record's similarity list postdate Ussery's filing. Under pre-AIA §102(a)/(b)/(e) as applied to a 1901 filing they are legally incapable of anticipating claims 1–6: US 765,504 (1904-07-19), US 807,849 (1905-12-19), US 923,872 (1909-06-08), US 935,596 (1909-09-28), US 942,762 (1909-12-07), US 996,635 (1911-07-04), US 1,038,051 (1912-09-10), US 1,050,064 (1913-01-07), US 1,051,145 (1913-01-21), US 1,138,455 (1915-05-04), US 1,144,241 (1915-06-22), US 1,177,453 (1916-03-28), US 1,219,191 (1917-03-13), US 1,302,562 (1919-05-06).

One flagged exception to that flat rule: a post-filing US patent can still be §102(e) art if its underlying application was filed before 1901-08-14. The nearest-miss candidates deserving a filing-date pull are US 765,504 (granted 1904-07-19) and US 807,849 (granted 1905-12-19). Given early-1900s pendency of roughly one to three years, both were probably filed after Ussery's date — but that is an inference, not verified fact.

---

Honest limitations

1. I found no citation to analyze, and I will not manufacture one. Every entry above is drawn from the record's machine-generated similarity list or from my own independent search, and each is labeled as such.
2. I retrieved full text for only three references (US 463,741, US 680,041, US 334,921) plus partial text for US 591,731. For the other nine pre-filing entries I have title and date only. No claim-by-claim §102 mapping is defensible for those nine without their specifications.
3. The "Similar Documents" list is not a search report. For a 1901 patent it is almost certainly a classification/keyword-overlap product (hence the "prior art keywords" bag, mail, arm, rod, cranes), which is why it mixes pre- and post-filing art indiscriminately.
4. US 447,471 and US 463,158 have corrupted title fields in the source ("John f") and were not assessable.
5. I could not complete a direct USPTO Patent Center / Patent Public Search full-text pull of 688,903 within this session; the analysis rests on the authoritative full text supplied plus web corroboration.

Recommended next steps

• Pull the filing dates for US 765,504 and US 807,849 to close out the §102(e) question.
• Pull the full specifications of US 501,150, US 511,509, US 471,270, US 481,673, US 655,669, US 195,567 — the six pre-1900 references whose titles most closely track claims 1–2.
• Verify US 591,731's grant date before citing it anywhere.
• If a §102 chart is required, the strongest single reference is US 680,041 (Harriott) — §102(a) and §102(e) art landing eight days before Ussery's filing — and the decisive question will be whether its pivoted bar d + spring-sliding rod e discloses claim 1's requirement of lateral pivotal movement on a vertically-yieldable axis.

Primary source: Google Patents US688903 (full text and structured record, fetched 2026-09-26).

I'll ground this § 103 analysis in the actual prior-art references surfaced on the patent's own page and verify what those references disclose.

Excellent — US 481,673 (Ray) is a critical find. Let me pull the remaining listed references to verify what each discloses.

Let me pull the remaining listed references directly.

Obviousness Analysis — U.S. Patent No. 688,903 (Ussery, "Bag-Holder for Mail-Bag Cranes")

Analysis date: April 26, 2026. Built on the litigation and PTAB sections already generated; those sections' conclusion (no litigation, no AIA proceeding, term expired 1918-12-17) is carried forward and not repeated.

---

I. Two threshold caveats before the merits

(1) The statutory vehicle is anachronistic, but the analysis is still meaningful. U.S. 688,903 was filed 1901-08-14 and granted 1901-12-17. 35 U.S.C. § 103 did not exist until the Patent Act of 1952. In 1901 the operative requirement was the "invention" condition of R.S. § 4886, construed in Hotchkiss v. Greenwood, 52 U.S. (11 How.) 248 (1851), and the crowded-art narrowing principle of Atlantic Works v. Brady, 107 U.S. 192 (1883). The modern Graham v. John Deere Co., 383 U.S. 1 (1966), and KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), framework maps onto the 1901 art cleanly, because Graham itself applied §§ 102/103 reasoning retrospectively to pre-1952 patents. So a § 103-style analysis is coherent — but the correct doctrinal label for a 1901 filing is "lack of invention," and the § 103 labels below are functional analogues, not historical claims of law. I flag this rather than silently applying § 103 as if it applied in 1901.

(2) The bottom line is academic. Because the term expired 1918-12-17, no claim can be adjudicated invalid or invalidated. This matters practically for only two scenarios: (a) an invalidity/validity opinion where 688,903 is asserted as prior art against a hypothetical later claim (possible in principle; the Ussery disclosure as a whole dates to the 1901-12-17 grant); or (b) a diligence memo requiring a hypothetical validity assessment as of 1901.

Sourcing note — what I verified and what I did not. The Google Patents page lists prior art only as titles/dates ("Similar Documents") plus four keywords (bag, mail, arm, rod, cranes). I verified the substance of US 463,741, US 481,673, and US 680,041 by retrieving their text. For US 471,270, US 511,509, US 501,150, US 655,669, US 195,567, US 447,471, US 463,158 and US 614,756, I have only the title and date; I did not verify their disclosures and will not describe them. Misdescribing an unread reference is the single easiest way to produce a defensible-looking but wrong § 103 chart, so those entries appear below only as "identified, not read."

---

II. The prior art set

A. Verified references (read)

B. Identified from the page's prior-art list but not read — no disclosure asserted

US 471,270 (Mail-bag hanger, 1892-03-22); US 511,509 (Mail-bag holder, 1893-12-26); US 501,150 (Automatically-folding crane for mail-bags, 1893-07-11); US 655,669 (Railway-mail crane, 1900-08-07); US 195,567 (Improvement in mail-bag cranes, 1877-09-25); US 447,471 (1891-03-03); US 463,158 (1891-11-17); US 614,756 (1898-11-22). These would need to be pulled in full to be cited in any chart. US 655,669 (1900) is the highest-value unread reference — a railway mail-crane granted a year before filing is a natural primary or secondary reference, and the fact that the page's own clustering algorithm put it in the "Similar Documents" set is meaningful but not a disclosure finding.

C. One reference I cannot responsibly use

My search surfaced a PDF at patentimages.storage.googleapis.com/.../US497391.pdf (1893-era subject matter: a crane clip whose barrel 30 receives a stud 31 and a pin 32 turning at right angles in a recess, carrying fingers 33 and 34, the longer finger being the one the ring is hung on, "so there is no liability of the bag being caught and the ring broken when it is removed"). The patent number in the URL was not confirmed against the document's own heading in the excerpt I retrieved, and I did not verify title, date, or claims. It is offered here only as a research lead. Also surfaced: US 1,016,478, US 885,404/US 835,404 (the search snippet is internally inconsistent on the number), and US 942,762 — all post-date 1901 and are therefore not prior art to 688,903 under any theory. Excluded.

D. The patent's own admissions (usable as prior art)

The Ussery specification concedes: the crane with post 1 and pivoted lower/upper arms 2, 3 "pivotally connected with the post in any usual or preferred manner"; and an arm "having its outer end reduced to present a finger 7 to engage with the ring 8 of the mailbag 9, as usual." Those two elements therefore drop out of the novelty/invention inquiry entirely.

---

III. Claim-by-claim obviousness

Claim 1 — "mail-bag engager mounted for vertical swinging movement upon a fixed transverse axis and for lateral pivotal movement on a vertically-yieldable axis"

Primary reference: Ray 481,673.

This is the closest reference in the entire field to claim 1, and the correspondence is not accidental. Ray's stated object is dual-axis yield — "yield vertically and laterally." Ray's structure supplies the crane-arm-plus-pivoted-gravity-arms environment (the same as admitted), a plate/bar C attached at the arm end, and a trigger journalled on it for two distinct degrees of freedom, with lateral-movement limiting means shown in Ray's Fig. 4.

Secondary: Scientific American 1897 supplies the express two-axis teaching in the same apparatus class — a catch arm on the crane whose universal coupling "permits it to swing both horizontally and vertically," with the coupling also performing the blow-cushioning function.

Motivation (KSR rationales, all grounded in the references themselves, not in hindsight):

• Same field, same problem, same solution mechanism. Ray and the 1897 article address the identical problem Ussery recites — rings/straps breaking on detachment, and unreliable release at speed.
• Identified, finite set of known options. Ray frames the art as: triggers yielding laterally only (old) → triggers yielding vertically and laterally (Ray). Adding the second orthogonal degree of freedom is the next step in a two-step progression the art itself lays out.
• Predictable result. Two-axis yield → the engager rotates out of the ring's path under a lateral/oblique pull → less resistance to release. No new principle of operation is required.

Honest weakness in the chart. Ray's "lateral" yield is described in terms of the triggers yielding toward each other (spring-relieved), and its journal c/bearing c′ may be a fixed pivot rather than a vertically-yieldable one. Ussery's claim 1 requires the lateral pivot axis itself to be vertically yieldable (spring 19 on rod 14 through plate opening 17). If Ray's lateral pivot is fixed and only its vertical adjustment is sprung, Ray alone does not literally meet claim 1 — but Ray plus the 1897 universal-coupling teaching plus the admitted spring-retained laterally-yielding trigger of Ray's own background section makes claim 1 obvious, because the springs-on-triggers arrangement was expressly admitted old and the substitution of a spring-loaded vertical pivot for Ray's vertical yield is a predictable mechanical variation. I regard claim 1 as obvious (high confidence) and as arguably anticipated by Ray (medium confidence — I did not retrieve Ray's claims, only its specification).

Claim 2 — claim 1 plus "spring-retracted axis"

Combination: Ray 481,673 + [Ray's admitted prior art: springs on trigger-carrying arms] + Greene 463,741 (wind-guard spring B) + Harriott 680,041 (expansive spring f biasing a pivoted bar).

Spring-retraction of a bag-engaging member is the most thoroughly anticipated element in this patent. Ray's own background section states that in the old laterally-yielding cranes the triggers were "held in proper position by means of springs secured to the arms carrying the triggers." Greene's wind-guard spring B exists precisely to prevent accidental disengagement. Harriott's spring f exists to throw the pivoted U-bar to its normal position. A spring performing automatic return-to-normal on a pivotally mounted bag holder is old, disclosed in three independent references, and admitted in the specification's own description of the art. Ray additionally supplies the motivation for spring force calibration — it notes springs "quickly become weakened," which is exactly the problem Ussery's adjustable nut (claim 6) addresses.

Claim 2 is obvious (high confidence); on the admitted prior art alone it may be anticipated.

Claim 3 — finger at one end, cam-bearing head at the opposite end, spring-pressed rod connected with the head

Combination: Greene 463,741 (pivoted holder: ring-engaging catch at one end, mounting structure at the other) + Harriott 680,041 (spring-pressed rod connected to and acting on a pivoted holder head, biased to normal, released by pivoting) + conventional cam-and-follower expedient.

Harriott is the key. Harriott's arm c′ / U-bar d / rod e / spring f / abutments c², e³ assembly is a spring-pressed rod connected with a pivoted holder head, held in a normal, bag-retaining position by spring force and thrown to a released position by the catcher's oblique pull — the same structural and functional relationship claim 3 recites. Harriott's aim is verbatim Ussery's aim: secure hold, certain release, no broken rings.

The "cam-bearing head" limitation is the weakest link in the claim, and I must be candid: I did not verify a pre-1901 reference expressly disclosing a cam surface on a bag-engager head bearing on a plate to effect automatic return. The argument is therefore a KSR "known technique / predictable variation" argument:

• A cam converting a spring's stored force into a restoring couple is one of the oldest mechanical expedients; the patentee himself describes surface 12 purely functionally ("to effect automatic return of the arm to its normal position"), meaning the spring does the work and the cam only guides it.
• The "flat bearing-surface 11 to engage with the plate 4" is a mere abutment, functionally identical to Harriott's abutments c²/e³ and Greene's groove-and-stop b/h.
• Where the only difference is the shape of a force-transmitting surface on an otherwise fully disclosed spring-and-pivot assembly, and the shape produces the result the artisan already wanted (return to normal), the variation is within the skill of the art.

Claim 3: obvious (medium-high confidence), with the cam-surface element carrying the burden.

Claim 4 — claim 3 plus "a stop or shoulder adjacent" the finger

Combination: claim 3's combination + Greene 463,741 (catch c with vertical part i, plus wind-guard spring) + the 1893-era two-finger clip (33/34, longer ring finger with adjacent inner finger "keeping the ring properly positioned" — unverified reference, flagged above) + Harriott's abutments.

Withholding nothing: the shoulder is a ring-retainer. The field was saturated with ring retainers — a wind-guard spring (Greene), two fingers on a pin with the inner one positioning the ring (the flagged 1893 clip), an abutment pair bracketing the rod (Harriott). Greene also uses a stop/shoulder (h) as a functional positioning element. Motivation: prevent the ring sliding off the finger and keep the ring presented at the right attitude for the catcher — stated in each of these references.

Claim 4: obvious (high confidence) on Greene + Harriott; a shoulder/stop is a de minimis mechanical addition.

Claim 5 — claim 3 plus "vertically-slotted" cam-bearing head, rod working in the slot

Combination: Harriott 680,041 (rod e sliding in the parallel members of U-shaped bar d — a slot-and-rod coupling, with pin e′ and spring f) + Greene 463,741 (grooved arm with stop b) + claim 3's combination.

Harriott's rod-in-U-channel is the functional equivalent of Ussery's rod-in-vertical-slot; both constrain a spring-loaded rod to a single line of travel while permitting the head to pivot. Greene's bolt-and-groove arrangement is a second instance of a constrained, stop-limited sliding coupling in the same apparatus.

Additional KSR "design choice" point that is unusually strong here: the Ussery specification itself states that the invention "is not to be limited to the precise manner of associating the rod 14 with the pin 16," and expressly offers the alternative of "a yoke having its terminals formed into hooks, the yoke to straddle the head and to engage the ends of the pin 16." A patentee who tells the world that the slotted head and the straddling hooked yoke are interchangeable alternatives has conceded that the slot is not the inventive contribution. An element the patentee characterizes as an arbitrary equivalent cannot supply the "invention" that sustains a claim to it.

Claim 5: obvious (high confidence).

Claim 6 — the full combination: plate; finger with shoulder; vertically-slotted cam-bearing head carrying a transverse pin; rod hooked at its upper end, lower end projecting beyond the plate; spring on the rod bearing against the plate; spring-tension-regulating nut

Claim 6 is a claim to an aggregation of individually old elements, and its elements map as follows:

The specific "spring-tension-regulating nut on the threaded rod" sub-element is the one for which I again have no verified pre-1901 reference showing a nut used specifically to adjust spring preload in this apparatus. I will not manufacture one. The argument is: (i) the nut-on-threaded-member convention in this exact assembly is shown by Greene; (ii) adjustable spring preload is a routine desideratum where bag weights vary, and the specification expressly frames the nut as an adjustment ("By turning the nut upon the rod let the tension of the spring may be adjusted to cause it to suspend a bag of any weight") — i.e., a tuning parameter, not a new mechanism; (iii) the claimed tension setting ("just sufficient tension to … resist the tendency to be tripped") is a result-effective optimization, not a structural contribution.

Claim 6: obvious (medium-high confidence) as an unpatentable aggregation; the strongest counter is the absence of a verified cam and a verified preload-nut reference, both of which I have flagged rather than papered over.

---

IV. Combination summary and the motivation narrative

Cross-cutting motivations that a § 103 analysis should record (all quoted or paraphrased from the references themselves, so no hindsight is imported):

1. The problem was stated identically by Ray (1892), Harriott (1901), and Ussery (1901): release the bag certainly when the catcher strikes, without tearing the bag or breaking the rings. Same problem → motivation to look to the same solutions.
2. The art was crowded and old. The 1897 Scientific American piece states the subject had "been of much interest to inventors through many years, and in relation to which many patents have been issued." Ray's background section recites a whole prior generation of laterally-yielding, spring-held triggers. Under Atlantic Works and the crowded-art narrowing principle — echoed verbatim in a period Circuit opinion at 73 F. 684 ("the field of invention was a very narrow one, and complainant's claim can be sustained only under a construction which will restrict it closely to the precise combination"), which I retrieved — the quantum of ingenuity needed to sustain these claims is high.
3. Ussery's own specification concedes the peripheries: "as usual" finger-and-ring engagement; crane arms "pivotally connected … in any usual or preferred manner"; and the slot-versus-yoke interchangeability. Concessions of this kind are the classic basis for an obviousness holding, because the patentee has defined the outer boundary of the alleged invention as conventional.
4. No secondary considerations are available. There is no evidence of commercial success, long-felt need, failure of others, or copying in the record, and none could now be generated: the last possible infringing act under the patent occurred in 1918. Where a patent owner offers no objective indicia, KSR directs that the obviousness inquiry rests on the references and the ordinary skill level.

---

V. Where a defender of the claims could push back

• Ray's axes are not Ussery's axes. Ray's two yields are (a) vertical and (b) triggers yielding toward each other; Ussery's are (a) vertical swing about pin 16 and (b) lateral rotation of the engager to a position at right angles to normal about rod 14. Ussery's second axis does more than relieve tension — it reorients the engager out of the ring's path. If that reorientation function is treated as the point of novelty, the 1897 universal-coupling disclosure (horizontal and vertical swinging of the catch arm) becomes essential to the chart, and I have only its abstracted journal description, not its full text or figures.
• The cam surface, the preload nut, and the vertically-yieldable (as opposed to merely fixed) lateral pivot are the three limitations for which I could not produce a verified pre-1901 disclosure. A rigorous chart should obtain: full text of Ray 481,673 including its claims; full text of US 655,669 (Railway-mail crane, 1900-08-07) and US 501,150 (1893-07-11), which are the likeliest homes for a cam-return and an adjustable-spring disclosure; and the verified identity of the US497391 reference.
• The examiner allowed these claims in 1901. That is entitled to no weight as precedent, but it does mean the Primary Examiner at the time thought Ray was not on point. That view predates KSR by 106 years and is not binding — but it is the honest counter-argument, and it is worth one sentence in any memo.

---

VI. Conclusion

Claims 1, 2, 4, and 5 of U.S. 688,903 are, on the verified record, obvious (and claims 1–2 are arguably anticipated by Ray 481,673, which I read in specification but not in claim form). Claims 3 and 6 are obvious to a medium-high degree of confidence, subject to two explicitly identified evidentiary gaps — a pre-1901 cam-return disclosure and a pre-1901 adjustable-spring-preload disclosure — neither of which I have verified and neither of which I will assert. The central technical insight of the patent, two-axis mounting of a bag engager on a mail-crane arm to guarantee release without ring damage, was stated as an object and reduced to structure by Ray in 1892 and by the 1897 Scientific American description, and the spring-pressed, pivoted, position-bracketed holder that release claims 3–6 was disclosed by Harriott in a patent granted eleven days before Ussery filed.

The correct 1901-era statement is therefore that the claims, measured against the art of 1877–1901, do not rise to invention; the correct modern statement is that they would not have been patentable under § 103. Both statements are, however, moot: the patent expired 1918-12-17, and no tribunal has ever or can ever rule on them.

Recommended documents to obtain before this analysis is filed anywhere: (1) full text and claims of US 481,673 (Ray); (2) US 655,669 (1900) full text; (3) US 501,150 (1893) full text; (4) US 511,509 (1893) full text; (5) verified identification of the US497391 reference; (6) US 471,270 (1892) and US 195,567 (1877) full texts. Items (1), (2), and (3) are, in my assessment, the ones most likely to close the remaining gaps on claims 3 and 6.