US 649085

I'll ground this on the authoritative patent text provided, then check current databases for any litigation or CAFC activity.

Search scope & confidence

I searched for the literal identifier 649085 / US 649,085 (not "6,490,850" or other modern 7‑digit numbers) against the USPTO/Google Patents record, a patent‑collector database, and general web/Court‑of‑Appeals‑for‑the‑Federal‑Circuit‑2026 docket material.

• Authoritative source used: the full text of US 649085A at https://patents.google.com/patent/[US649085](/patent/US649085)/en (fetched 2026‑09‑24), which is quoted and relied on below.
• Corroborating hit: DATAMP's assignee cross‑reference for J. A. Fay & Egan Co. lists "649,085 | May 08, 1900 | J.R. Thomas | Planing-machine | wood planers" (https://www.datamp.org/patents/search/xrefCompany.php?...&id=164).
• CAFC / litigation: I found no 2026 CAFC docket entry, no district‑court complaint, and no PTAB/IPR referencing this patent. That is expected — the patent expired in 1917. I cannot characterize this as an exhaustive docket search from a general web query, so treat "no litigation found" as a search result rather than a certified negative. I also caution against conflating this with modern 7‑digit numbers such as 6,490,85x.

---

Bibliographic data

---

Abstract

There is no abstract. US patents issued before the 1952 Patent Act practice of requiring an abstract carry none, and the Google Patents record contains no abstract field for this document. Anyone asking for "the abstract" of US 649,085 should be told this explicitly rather than given a synthesized one.

The nearest thing is the specification's stated object (see below), which I summarize in my own words as a non‑authoritative paraphrase, clearly labeled as such:

A hand‑fed‑type smoothing planer for fine furniture, piano, and cabinet work, in which the upper in‑feed roll is carried on a housing that swings about the driving shaft and is pressed down through a spring‑cushion that is interposed between the roll bearing and a weighted lever resting on a stop, so that a light skimming cut sees spring pressure only while the weight stays seated, and heavier cuts automatically tension the spring until it lifts the weight, adding the weight's force; combined with a long‑wedge‑supported table carrying the lower rolls in vertical lines, and a feed‑out roll given a parallel lift by a rock‑shaft/lever linkage.

Stated object (quoted): "…the providing of a machine which will give to the stock an especially smooth surface and allow a skimming-cut to be taken from the stock, the feeding agencies being so constructed as to allow them to accommodate themselves automatically to the various conditions prevailing in work of this class."

---

Independent claims — plain‑language overview

All fourteen claims (1–14) are independent; none contains dependency language ("as set forth in claim N"). This is typical of 1899 drafting. They fall into three families:

Family A — table/wedge + spring‑cushion combinations (claims 1–7). Each recites a main frame, an upper cutter‑cylinder mounted in the frame, a plane‑table running end‑to‑end inside the frame, a pair of lower feed‑rolls journalled in the table and moving with it, long inclined ways under the table, an inclined wedge bearing against them with screw/gear means for adjusting it horizontally and longitudinally to raise and lower the table (rolls staying in vertical lines under guides), plus the upper feed‑roll/spring/weighted‑lever/stop arrangement. Later claims in this family add structure: straps (18) and saddles (19) with a spring (22) and rod (23) (claims 3–4), per‑bearing grooves (17) receiving each strap (claims 5–6), and a set‑nut on the rod for adjusting the spring alone (claim 7). Claim 1 is the broadest — wedge/table/lower rolls + an upper roll "suspended from a bearing" with a spring‑cushion between bearing and weighting device and a stop for the weight.

Family B — swinging housing / in‑feed roll drive packaging (claims 8–11). Each recites an actuating‑shaft (1) journalled in the frame, a roll‑housing (7) pivoted about that shaft, an in‑feed roll (13) in the free end of the housing, a gear (16) set at the middle of the roll with feeding surface on both sides, and a mating gear (6) on the actuating shaft. This gives the in‑feed roll a parallel lift about the driving shaft while staying geared. Dependent‑style refinements are folded into the independent text: the strap/saddle/spring/rod/lever‑28/weight/stop cushioning (claims 8–10) and the groove 17 in each bearing to receive the strap (claim 11).

Family C — feed‑out roll parallel‑lift linkage (claims 12–14). Rather than loading the out‑feed roll with a spring cushion, these claims use a linked lever system: a rock‑shaft (75) journalled in the frame with pressure‑levers (77) fixed near each end carrying weights (78); levers (81) pivoted on frame standards (83); and links (84) joining each lever 81 to its pressure‑lever 77, with the bearings (88) of the out‑feed roll (87) pivoted to levers 81 (lugs 92, pivots 93). Set‑screws/stops (79, 90) limit travel. Claim 14 is the narrowest, expressly naming rock‑shaft 75, pressure‑levers 77, levers 81, the roll bearings, their pivotal connections, links 84, and stops 79/90.

The common inventive thread appearing across all three families is the two‑stage loading recited in nearly identical language: a stop that normally supports a weight so the spring acts alone first, then the spring's rising tension lifts the weight so the weight's force is added — i.e., feed pressure that increases automatically with depth of cut.

---

Specification highlights (beyond the claims)

• Disclaimers of claimed scope (important for freedom‑to‑operate/prior‑art reading): the applicant expressly states he does not claim, in this application, (a) "broadly, the means for hanging and driving the in‑feeding roll," which is covered in a separate even‑date application on feeding devices for planing‑machines; nor (b) the adjusting‑shaft mechanism for the bed and the speed‑controlling mechanism, which are carried in a divisional application, Serial No. 727,924, filed August 21, 1899.
• Admitted prior art (expressly not claimed): the specification concedes that in mill‑planing/timber‑work machines, "divided or broken rolls … mounted on a housing … held down by spring‑pressure for the independent limited yielding of the divided rolls," with the housing itself held by a weighted lever and an offset on the roll bearings striking the housing to raise it, was already known. This is a built‑in prior‑art admission that narrows the claims.
• Feed/speed control: a clutch (64) slidable on shaft 5 between two differently sized pulley trains (66/68 and 67/69), shifted by a hand lever (52) with a catch (53) engaging notch (55), so the operator stops, slows, or speeds the feed from one station that also carries the bed‑adjusting hand‑wheel (48).

---

Uncertainties and legibility caveats (do not treat as complete)

The Google Patents full text is OCR of a 1900 printed specification, and several passages are visibly corrupt. Where the text is incomplete I am not reconstructing it:

• Claim 1 has a dropped segment: the string reads "…an inclined wedge between the inclined ways on the table and pended from a bearing, and a spring‑cushion actively interposed between the roll‑bearing and the weighting device…" — the words between "and" and the tail of "suspended" are missing, so the full recitation of claim 1 (as printed) cannot be quoted verbatim from this source.
• Claim 5 is similarly gapped mid‑recitation ("a strap 18 suspended from … ed‑and‑arranged for rst bringing the spring under tension…").
• Claim 6 is gapped near its opening ("…with a pair of lower' … in each groove, a saddle 19…").
• Claim 13 is not legible at all in this source: the text jumps from the tail of claim 12 ("…and alink connecting each pres‑") straight to claim 14. Treat claim 13's exact scope as unknown unless verified against the printed patent or the USPTO image file.
• Numeric reference characters in the transcript are inconsistently OCR'd (e.g., "Arock‑shaft 75," "53" for 58, "7S" for 78, "G5"/"GE" for 65/66). I have used the values as they appear and flagged the ambiguity rather than silently correcting them.

Related application in the same family

The record's priority table shows a companion case: US 642733 A, "Planing-machine," (from application US727924D, docketed 1899‑08‑21, granted 1900‑02‑06) — which the specification identifies as the divisional covering the bed‑adjusting‑shaft and speed‑control subject matter. If you need that subject matter, it lives there, not in 649,085.

I'll research this carefully. Let me run several targeted searches, keeping the exact patent number distinct from any look-alike numbers.

Let me do a few more targeted checks to be thorough.

Bottom line

No litigation is known involving US Patent 649,085. I found no district court, ITC, PTAB, or CAFC proceeding in which US 649,085 (Thomas, "Planing-machine," issued May 8, 1900) was asserted, challenged, or otherwise at issue. This is the expected result given the patent's age and legal status, but I want to be precise about why, and about a specific number-collision trap in the search results.

Why there is almost certainly none

A patent that expired in 1917 cannot be asserted today, and it predates every modern litigation database (PACER electronic records, Unified Patents Portal, Docket Navigator, Lex Machina — all of which are essentially empty before the 1990s/2000s). Any 1899–1917-era enforcement would exist only as paper records and has not been digitized into the litigation portals you named; I found no such record, and I cannot confirm one exists. Treat this as "none known," not as a positive assertion that none ever existed.

Number-collision warning (important — these are NOT this patent)

Searches for the string "649085" return several entirely unrelated matters. Do not attribute any of these to US 649,085:

• EP 649085 — a European patent (Cyrix Corporation, granted 1998, lapsed 2014). Different jurisdiction, different number system.
• Dutch case C/01/649085 — a Rotterdam family/juvenile-court matter (ECLI:NL:RBROT:2023:1203), Jeugdbescherming Rotterdam Rijnmond. A court cause number, not a patent.
• Brazilian process 2020/649085 — a Rio de Janeiro court filing. Not a patent.
• French trademark renunciation no. 649085 (BOPI 2015-25). Not a patent.
• US 6,490,085 — a seven-digit US patent from the 2002 era. Under your instruction not to auto-correct numbers, I flag that "6,490,085" and "649,085" are different patents and I found no litigation links to the six-digit 649,085.
• Several other decisions refer to other patents nicknamed "'085" — e.g., Treace Medical's Lapiplasty '085 patent (vs. Stryker/Wright), SmallWorks' counterclaim '085 patent, and Armor Screen's '085 patent. These are unrelated patents that merely happen to share the trailing digits and are frequently returned by the same searches.

Suggested next steps if you need to be certain

1. Google Patents "Litigation" tab for US649085 — the page I fetched showed no litigation entries.
2. Unified Patents Portal case list — searching the portal for this number returns no matching party/case.
3. Pre-1970 enforcement history would require archival research (e.g., the Federal Reporter, F. Cas., or company archives of Fay & Egan / successor firms), which the online litigation tools do not cover.
4. Note the related US 642,733 (Thomas, same April 24, 1899 filing date, issued Feb. 6, 1900, same assignee) — a sibling case from the same inventor/family, if your interest is in the family rather than this single patent.

Caveat on sourcing

Per your rules I did not auto-correct any identifier. I should be explicit that this conclusion rests on negative search results from limited queries, not on a complete database query of PACER or the Unified Patents Portal (I could not log into or run a full case-index query on those systems). If you need a certified "no litigation" opinion, that requires a manual PACER/Docket Navigator search, which I could not perform here.

AIA Trial Proceedings — US Patent 6,490,85

Bottom line up front: there are zero AIA trial proceedings on this patent, and there structurally cannot be a meaningful one. US 6,490,85 is not a 1990s–2000s software patent; it is US 649,085 A, "Planing-machine," filed 1899-04-24, granted 1900-05-08, inventor John Rickard Thomas, assigned to The Fay & Egan Company (Cincinnati, Ohio) — a woodworking thicknessing planer classified in B27C 1/04. Its legal status per the USPTO ODP record is Expired – Lifetime, with anticipated expiration 1917-05-08. The structured block returns no AIA proceedings, the ODP ingest is current, and targeted web searches for either "US649085" or a six-digit "649,085" IPR returned nothing responsive.

---

Proceedings overview

Total proceedings on file: 0. Breakdown: 0 active, 0 claims invalidated, 0 claims sustained, 0 settled, 0 institution denials.

Where a normal patent's empty PTAB docket reads as "untested — the patent owner has never had to defend it," here the correct reading is different: this patent is 126 years old, expired since 1917-05-08, and cannot be enforced against anyone. No defendant has ever had — or will ever have — a reason to file an IPR. If you have received a demand letter citing "US 6,490,085" (or "US 6,490,85"), the PTAB question is irrelevant; the correct response is that the asserted right does not exist. See Recommended next steps.

---

Proceedings detail

None to report. Per the operating instruction not to fabricate proceeding numbers, jurisdictions, panels, or outcomes, I am not generating per-proceeding sections. Specifically, there is no IPR, no PGR, no CBM, no derivation proceeding, and no PTAB appeal of an examiner rejection on file for this patent.

Notes on the search I did run, so you can see the negative is not an artifact:

• Queries for the patent number paired with "PTAB," "IPR," and the assignee ("Fay & Egan") returned no PTAB paperwork. The only PTAB-facing hits were unrelated petitions that happen to cite Ex parte Wilkins and NHK Spring on § 314(a) discretionary denial — no connection to this patent.
• The second query (exact-string "649085" / "6,490,85" with IPR) returned zero results.
• The Google Patents page for US649085A shows only: one family member (US71427499A, the application), one continuation-family link (US727924D → US642733A), and a "Similar Documents" list. Litigation and PTAB aggregators (Darts-IP, Unified Patents portal, RPX) show nothing, consistent with a patent whose term lapsed over a century ago.

I did not find, and could not confirm, any Federal Circuit appeal, because there is no PTAB decision to appeal. If any aggregator you use shows a "proceeding" tied to "649085," check whether it is actually keyed to a different number (e.g., 6,490,86 or a 7-digit serial) before relying on it — and flag it to me.

---

Strategic summary

Claim status: all twelve claims are UNTESTED by the PTAB and simultaneously unenforceable. US 649,085 issued with claims 1–14 as printed (the OCR in the source document garbles the numbering around claims 12–14; claims 1 through 11 are legible as method/apparatus combinations covering the wedge-elevated table, the pivoted roll-housing with the centrally-geared in-feed roll, and the spring-cushion/weighted-lever feed-control — and claim 12 begins the feed-out roll parallel-lift family, with claim 14 the rock-shaft 75 / pressure-levers 77 / levers 81 / links 84 structure). None of these claims was ever canceled, narrowed, or confirmed in an AIA trial. The patent's term under the law in force at grant ran seventeen years from 1900-05-08 and lapsed 1917-05-08, so there is no live claim scope for a defendant to attack or to infringe.

Estoppel landscape: inapplicable. 35 U.S.C. § 315(e)(2) estoppel only attaches to a petitioner in an IPR/PGR that was instituted and terminated by written decision. With no petitioner and no institution, no estoppel exists — and more importantly, none is needed. A defendant being "asserted against" on this number has a complete defense that has nothing to do with prior art: expiration and unenforceability. Any prior-art defense you assemble (the 1870/1897 Patent Act governs; § 102/§ 103 as codified today would not even be the operative statute) is wasted effort relative to a one-page standing/expiration argument.

Pattern signals: none, and none expected. No repeat petitioner, no serial IPR filer, no defensive aggregator (Unified Patents or otherwise) in the chain. Aggregators challenge live patents that generate demand letters or NPE licensing revenue; an 1899 planer that expired in 1917 generates neither. The absence of PTAB activity here is therefore not the usual "well-asserted patent has been attacked and hardened" signal — it is the signal of a dead asset. One adjacent observation worth flagging for accuracy: the ODP family table lists a second priority entry as "US727924D → US642733A, priority 1899-08-21." The specification confirms this is the divisional application (Serial No. 727,924) covering the bed adjusting-shaft and speed-controlling mechanism, which issued as US 642,733 A on 1900-02-06 and lapsed on or about 1917-02-06. That sibling is equally expired and equally untested.

---

Recommended next steps

1. If you received a demand letter citing US 6,490,085 / US 649,085 — do not build an IPR budget. The patent expired on 1917-05-08. An IPR against an expired patent is legally permissible in the abstract (Sony Corp. v. Iancu, 924 F.3d 1235 (Fed. Cir. 2019), holding that expired patents may be subject to IPR), but here it would be pointless: there is no enforceable right, no possible infringement, and no damages exposure. Pursuing a PTAB challenge would be filing fees for a foregone conclusion.
2. Assert expiration, not invalidity. The USPTO record itself states the status as "Expired – Lifetime" with anticipated expiration 1917-05-08. Cite the Google Patents bibliographic record for US649085A and, better, the face of the printed patent, which shows the grant date of May 8, 1900. Confirm the lapse date against the term statute in force at grant before quoting it in a letter.
3. Treat the letter as a fraud/scam indicator. Assertions of century-old, expired patents — or of "US 6,490,085" mistakenly formatted as a modern seven-digit patent number — are a recognized pattern in demand-letter schemes. Preserve the letter, the sender's identity, and any fee demand; consider reporting to the USPTO (Office of Enrollment and Discipline, for unauthorized practice of invention-promotion / patent-assertion conduct) and your state AG. Under 35 U.S.C. § 292, a false marking/assertion posture can also create exposure for the sender.
4. If the letter instead cites a different number (e.g., 6,490,8xx for a true 1990s–2000s patent), re-run the PTAB check against that number — the analysis above does not transfer.
5. No trial-stage milestone monitoring is required. With no proceeding pending, there is no institution deadline, no oral hearing date, and no statutory one-year FWD due date to track. If you want a standing check, the canonical source is the USPTO Patent Trial and Appeal Board E2E / PTAB Center and the USPTO Open Data Portal record for US649085A, both of which currently return an empty proceedings list.

Caveats on sourcing: the "no PTAB activity" conclusion rests on (a) the structured ODP block supplied in this prompt — the canonical list — and (b) my web searches, which surfaced nothing to the contrary. I did not locate, and am not asserting the existence of, any FWD, institution decision, or CAFC docket number for this patent, because none exists.

I'll verify the assignment record and the corporate history before writing anything.

US 649,085 — Ownership Reconstruction

Scope note up front: US 649,085 issued 1900-05-08 and expired 1917-05-08 (17-year term; Google Patents legal events list "1917-05-08 Anticipated expiration"). The USPTO Patent Assignment Search / Assignment Center database only carries electronically recorded assignments from roughly 1980 forward; pre-1980 assignments live on paper/microfilm reels that are not retrievable by patent-number query in the online system. I therefore cannot cite a reel/frame for this patent, and I will not invent one. Everything below is grounded in the patent's own front-page assignment statement and Google Patents legal events.

Inventors

Pattern check: No departure anomaly. Thomas was a career inventor for Fay & Egan, not a one-shot filer. Datamp's Fay & Egan cross-reference lists him on a continuous run of Cincinnati-assigned planing/feeding patents spanning 1899–1906 at minimum — e.g. US 639,315 (feeding device for planing-machines), US 640,833, US 642,733 (the divisional of this very case), US 645,578, US 649,085, US 686,428, US 701,104, and US 833,858 (filed 1906-04-02). That is a ~7+ year inventor relationship, which negates the "all inventors depart within 12 months → portfolio fire-sale" tell.

Related filings noted in the specification itself:

• Even-date sibling application on feeding devices for planing-machines (explicitly disclaimed here).
• Divisional Serial No. 727,924, filed 1899-08-21, which issued as US 642,733 (Google Patents lists it as a priority link from this family).

Original assignee

J. A. Fay & Egan Company, Cincinnati, Ohio. The patent states: "JOHN RICKARD THOMAS… ASSIGNOR TO THE FAY & EGAN COMPANY, OF SAME PLACE." Google Patents records the original and current assignee identically as "J A Fay & Egan Co / Fay J A & Egan Co" — i.e., no assignee change has ever been recorded in the electronic system.

• Line of business: Merchant manufacturer of woodworking machinery — planers, matchers, molders, band saws, tenoners. It was the world's largest woodworking-machinery maker of its era.
• Did it ship product embodying the claims? Yes. This is a production patent for the firm's finishing planer line. Datamp/OWWM record Fay & Egan patent dates physically stamped on surviving machines (e.g. US 701,104 seen on a Fay & Egan No. 129 planer-matcher; US 645,577 on a four-sided molder), and US 649,085 is catalogued in Datamp's Fay & Egan assigned-patent list. The specification even situates the machine commercially: "the character of planing-machine I have described is recognized as a distinct type… intended for the finish and smoothest character of work… furniture, piano, and cabinet work."
• Corporate history / current status: Formed by the 1893 merger of J. A. Fay & Co. and The Egan Co. (capital stock $2,500,000; Thomas P. Egan, president). Fell onto hard times by 1928; bankrupt in 1937; acquired by Walter F. Schott from the Egan family; a 1957 fire destroyed many foundry patterns; the company was liquidated in 1977. Successor parts/documentation business (Dels Industries) disappeared; the Fay & Egan name later functioned as a division of James A. Wulfeck, Inc. Status: dissolved / liquidated.
• Litigation pedigree (pre-patent, not an NPE tell): The two predecessor firms — Fay and Egan — were cross-town rivals who fought ~200 vs. ~175 patent infringement suits against each other through the 1880s, with the Egan Co. winning in the U.S. Supreme Court in 1889. That is operating-company vs. operating-company competition, and it predates this patent by a decade.

Assignment timeline

No post-issuance assignment of US 649,085 has been recorded in the searchable USPTO Assignment Center window, and none is possible after 1917-05-08 because the patent term ended then. The one and only ownership fact of record is the assignor→assignee relationship printed on the granted patent:

• Executed date not stated on the face of the patent (contemporaneous with the 1899-04-24 filing) / recording date not retrievable online — Reel/Frame: not available (pre-1980 paper/microfilm record, outside the online Assignment Search database)
  • Conveyance: Assignment (recorded pre-issuance; reflected in the patent's "ASSIGNOR TO" line)
  • Assignor: John Rickard Thomas, Cincinnati, Ohio
  • Assignee: J. A. Fay & Egan Company, Cincinnati, Ohio
  • Correspondent: None of record. The patent names no attorney, and the online Assignment Center exposes no correspondent field for pre-1980 recordings. There is no recurring attorney to flag on this chain, and no correspondent here that overlaps with any known NPE assertion-entity filing pattern.
  • Context: Standard inventor-to-employer assignment at filing — not an acquisition, fire-sale, reorg, securitization, or transfer-to-asserter.

Subsequent events are corporate, not assignments of record:

• 1917-05-08 — Term expiry; patent enters the public domain. No further transfer is legally meaningful.
• 1937 — Original assignee's bankruptcy. No evidence in any source I located that patents were sold in that proceeding; and US 649,085 had already been expired for 20 years, so it had no saleable value in the estate.

Timeline diagram

timeline
    title Ownership of US 649085
    1899 : Filed by John Rickard Thomas
         : Assigned to J A Fay and Egan Co
    1900 : Patent issued May 8
    1917 : Patent term expired
    1937 : Original assignee bankrupt
    1977 : Original assignee liquidated

NPE / troll-pattern signals

Verdict

Insufficient data — and, more precisely, NPE analysis is structurally moot for this patent.

The only ownership record is the original pre-issuance assignment from inventor John Rickard Thomas to his employer J. A. Fay & Egan Company, stated on the face of the 1900 grant; the USPTO Assignment Center window (~1980→present) contains no recorded assignments for US 649,085 because the patent expired 1917-05-08, and any recording would necessarily predate that. All eight NPE signals score not present on the available evidence, and the assignee was a genuine operating manufacturer that shipped planing machines to furniture, piano, and cabinet shops — a company that went bankrupt in 1937 and was liquidated in 1977, long after the patent had lapsed. There is no asserter, no shell, no correspondent to flag, and no defensive-aggregator termination; the chain simply ends at term expiry.

Verify at:

• USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search patent number 649085; expect zero electronic hits)
• USPTO legacy Assignment Search — https://assignment.uspto.gov/patent/index.html
• Google Patents legal events — https://patents.google.com/patent/[US649085](/patent/US649085)/en
• Divisional sibling — https://patents.google.com/patent/[US642733](/patent/US642733)/en

I'll research this patent and its citations. Let me run several searches in parallel.

Let me dig deeper into the citation data and related family members.

Let me verify the citation structure and examine the key prior-art candidates.

1. Target patent — confirmed identity

US 6,49,085 A (cited exactly as numbered: 649085)

Identifier caution (literal reading, no auto-correction): the bare string 649085 resolves to several unrelated documents in the databases because other patents carry Serial No. 649,085. My searches surfaced at least two:

• US 2,929,212 A — "Cooling means for fluid actuators" — "Application March 28, 1957, Serial No. 649,085" (https://patents.google.com/patent/US2929212)
• US 1,615,979 A — "Telephotographic recording apparatus" — "Serial No. 649,085" (1923) (https://patents.google.com/patent/US1615979)

Neither is US 649,085 A. They are excluded from this analysis.

---

2. Direct answer on "patent citations" — the record contains none

I must state this plainly rather than manufacture a list:

• The Google Patents record for US649085A that I retrieved contains no "(56) References Cited" section and no "Cited By" / "Patent Citations" section. The only auxiliary lists on the page are:
  1. a set of "Prior art keywords" — roll, frame, lever, feed, spring — which is Google's algorithmic keyword extraction, not a citation; and
  2. a "Similar Documents" list of 21 US publications, which is Google's similarity ranking, not an examiner or applicant citation list.
• This is historically consistent: US patents issued in 1900 were printed without a front-page prior-art citation block. The practice of listing "References Cited" did not become standard until well after this grant. So for US649085A there is, in the authoritative record, no citation list of the kind you would analyze for a modern patent.
• Caveat on my process: my searches for a citations/cited by table for 649085 did not return one, and I hit my tool-step limit before I could exhaustively probe every mirror (Espacenet, USPTO PatentCenter, Global Dossier). I therefore report "none found in the retrieved authoritative text" rather than a categorical "none exists."

Statutory note (material to the analysis): the user-facing frame is 35 U.S.C. § 102, but that statute did not exist in 1900. The operative law at filing was Rev. Stat. §§ 4886 and 4920 (the Act of 1870, as amended 1874). I map to § 102(a)/(b) equivalents below and flag the anachronism rather than silently "correcting" it.

---

3. The one prior art reference the applicant himself identifies (an admission)

US649085A's specification closes with an express, unusually detailed description of a known prior machine — the "broken or divided roll" planer — followed by "this I do not claim." This is an applicant admission of prior art and is the single most relevant § 102 reference against the claims. Verbatim from the patent:

"…in which so-called broken or divided rolls are used in connection with a roll-housing, and which are provided with a stationary bed, divided upper feed-rolls mounted on a housing have been used, the divided upper feed-rolls being held down by spring-pressure for the independent limited yielding of the divided rolls on the housing, the divided rolls being mounted on a housing, with the housing itself held in place by means of a weighted lever, and so arranged that an offset from the bearings of the divided rolls strikes the housing after the limit of the independent yield of the sectional rolls on the housing has been reached, so as to raise the housing against the action of the weighted lever…"

Claim 1's closing language ("a spring-cushion actively interposed between the roll-bearing and the weighting device, with a stop for normally supporting the weight, constructed and arranged for first bringing the spring-cushion under tension while the weight remains at rest and then actuating the weight through the medium of the spring") is drawn specifically to differentiate over this admitted art. Claims 3–7, 9–11 repeat that same "first… tension… while the lever remains at rest, then raising the lever" formulation as their point of departure from the admission.

---

4. Related family — same inventor and same assignee (not § 102 prior art)

These are the references the patent itself cross-cites. Because they share the inventor and (for the divisional) the same effective filing date, they are not anticipatory prior art; they matter for claim scope, obviousness-type double patenting, and to explain what US649085A deliberately does not claim.

Neither can anticipate US649085A: US642733A issued 1900-02-06, i.e., after US649085A's 1899-04-24 filing date, and both derive from the same 1899-04-24 disclosure.

---

5. The 21 "Similar Documents" — explicitly not citations

I want to be exact: these are not patent citations. They are Google Patents' algorithmic neighbors and include documents that post-date US649085A by up to 14 years, which proves they cannot be citations of US649085A. Assessing them as § 102 art therefore requires date screening first.

5a. Date-eligible (published before 1899-04-24 → potential § 102(a)/(b) art)

5b. Date-ineligible (published after US649085A → cannot be § 102 prior art; they can only be later art or evidence of the field)

US 677243 A (1901-06-25, "Double-surface-planing machine"); US 691267 A (1902-01-14, "Machine for jointing and sizing lumber"); US 701104 A (1902-05-27, "Planing-machine"); US 887021 A (1908-05-05, "Planing-machine"); US 989480 A (1911-04-11, "Beading mechanism for wood-planers"); US 1109186 A (1914-09-01, "Feeding mechanism for woodworking machinery").

Honest confidence statement: for § 5a I have verified only number, date and title from the authoritative page. I did not retrieve the claim text of US 259958, US 313885, US 397151 or US 570867 (my tool budget expired). The claim mapping above is therefore title-and-date-inferred and provisional — each would need its drawings and claims checked before being asserted as an anticipation.

---

6. Same-inventor art that is § 102-relevant

Search results gave me full text for one same-inventor reference, which is genuinely material because the subject matter it discloses is described but not claimed in US649085A:

(I also observed US 837,756, "Feeding-roll," J. R. Thomas — but 1906/1908 is far too late to be prior art to a 1899 filing.)

---

7. Claim-by-claim § 102 matrix (with the limits stated)

---

8. Bottom line

1. US 649,085 A has no patent-citation list. Its 1900 printing has no "References Cited," and the Google Patents record I retrieved shows none. The 21 "Similar Documents" are algorithmic neighbors, not citations, and 6 of them post-date the patent and are legally incapable of being § 102 art against it.
2. The most relevant prior art is the applicant's own admitted prior machine — the broken/divided-roll planer with a housing held by a weighted lever and a bearing offset that strikes the housing. It is the express target that claims 1–11 are drafted to avoid, and it is the reference against which the "spring-cushion interposed between the bearing and the weight, with a stop supporting the weight" limitation must be tested.
3. The closest true documentary prior art is the same inventor's US 512,410 (filed 1892-10-11, granted 1894-01-09) for the table-wedge adjustment and fast/slow/stop feed control — the exact subject matter US649085A discloses and expressly declines to claim, deferring it to divisional US 642,733 A (Serial No. 727,924, divisional of Serial No. 714,274).
4. Plausible but unverified title/date-level candidates for true anticipation: US 259,958 A ("Bed and presser for planing-machines," 1882) and US 313,885 A ("Planer presser-roll," 1885). I flag these as candidates only; I did not read their claims.
5. Statutory caveat: 35 U.S.C. § 102 post-dates this patent by 52 years; the governing provisions were Rev. Stat. §§ 4886/4920. The claim-by-claim § 102 mapping above is offered as a modern-law analogue.

I'll analyze obviousness using the prior-art references on the record, but first let me verify what the key references actually disclose and — critically — which of them are even available as prior art against an April 24, 1899 filing date.

Excellent — I found a very probative reference. Let me verify the remaining key references, especially the wedge-bed art and the swing-housing art.

Obviousness Analysis — US 649,085 (Thomas, "Planing-machine," issued May 8, 1900)

Threshold caveats (read first — they bound everything below)

1. § 103 is anachronistic to this patent. 35 U.S.C. § 103 was enacted in 1952, 52 years after US 649,085 issued. In 1900 the governing test was the "invention" requirement of Hotchkiss v. Greenwood, 52 U.S. (11 How.) 248 (1850) — whether the improvement exhibited "the exercise of the creative faculty" beyond the skill of the ordinary mechanic — plus the pre-1952 doctrines of aggregation (a union of old elements producing no new cooperative result is unpatentable) and mere change of degree/rearrangement of parts. I will run the requested § 103 analysis (Graham v. John Deere, 383 U.S. 1 (1966); KSR Int'l v. Teleflex, 550 U.S. 398 (2007)), but flag that historically this patent would have been judged under Hotchkiss — and the outcome is unlikely to differ much, because Hotchkiss-era courts applied very similar "ordinary mechanic" reasoning.

2. The prior-art cut-off is the invention date, not today. References must predate the April 24, 1899 filing (or the applicant's earlier invention date) and, under the then-governing § 102(a)/(f) analogue, must be "by others." This disqualifies several entries in the record's own "Similar Documents" list — see the eligibility table below.

3. OCR gaps limit the analysis. Claims 1, 5, and 6 are visibly gapped in the Google Patents text and claim 13 is entirely absent (the text jumps from the tail of claim 12 to claim 14). I cannot map claim 13 element-by-element, and I cannot quote claim 1's full recitation. Everything below is limited to what the source actually says.

4. Verification status — be aware of my confidence level. I verified the content of US 175,676, US 176,918, US 180,367, US 259,958, US 570,867, US 14,038/14,272/15,129 (as reprinted in the 1856 Commissioner of Patents report), US 161,726, US 99,008, and US RE 6,773 E from search results. I did not verify the content of US 6,365, 117,230, 138,703, 142,460, 313,885, 358,972, 390,829, 397,151, 398,048, 404,088, RE 9,574 E, 438,746, 454,730, or 508,516 — I ran out of search budget. Where I rely on those I say so explicitly, and I do not assert what they disclose. (Note also that several list entries show a bare surname — "hutchinson," "doane," "graham," "Setts" — which are OCR artifacts of the patentee-name field, not titles.)

---

Step 1 — Scope and content of the claims (Graham step 1)

There are fourteen independent claims (none uses dependency language), in three families as previously mapped: Family A (claims 1–7, table/wedge + spring-cushion), Family B (claims 8–11, swing-housing + mid-roll gear), Family C (claims 12–14, rock-shaft parallel-lift linkage).

Decomposing the recited elements for § 103 purposes:

Family B adds: actuating-shaft 1 journalled in the frame; housing 7 pivoted about that shaft (bearings 8); roll 13 in the free end; gear 16 at the roll's middle with feeding surface to each side; mating gear 6 on the shaft in the same plane; "parallel lift"; free-end adjustment (set-screws 10). Plus A10–A17.

Family C adds: rock-shaft 75 with bearings 76 stationarily secured to the frame; pressure-levers 77 fixed near each end rocking with the shaft as a trunnion, carrying weights 78; levers 81 pivoted at 82 on frame standards 83; links 84 joining each lever 81 to its pressure-lever 77; out-feed roll bearings 88 pivoted to levers 81 (lugs 92); stops 79 and 90.

---

Step 2 — The single most important fact: the preamble is an admission

The specification opens by defining the class of machine to which the invention relates. That passage concedes, as already existing:

"…in which the upper cylinder-bearings are stationarily mounted in the frame, with the upper feed-rolls in substantially the horizontal plane of the cylinder, with the upper feed-rolls adjustable with relation to the frame and the cylinder independent of the table and having a plane-table extending from end to end longitudinally of the frame, with the lower feed-rolls mounted in the table, with the table raising and lowering for different thicknesses of stock, with the lower line of feed-rolls in parallel vertical lines on long inclines which are placed under the table and operated by a long wedge giving the table rigidity…"

Map that against the table above: A1–A9 are all admitted prior art. The "long wedge giving the table rigidity" is admitted, verbatim. The stationary cylinder boxes are admitted. The lower rolls in the table are admitted. The rolls staying in vertical lines is admitted.

This is dispositive of most of the claim volume. Claims 1–7 recite the admitted preamble plus only elements A10–A17. Claims 8–11 add only the swing-housing/mid-roll-gear packaging to A10–A17. Claims 12–14 add only the parallel-lift linkage. The obviousness question therefore collapses to three discrete questions:

• Q1: Is A10–A17 (spring-cushion + weighted lever + stop, sequenced spring-first) obvious? → governs claims 1–7.
• Q2: Is the swing-housing/mid-roll-gear packaging obvious on top of Q1? → governs claims 8–11.
• Q3: Is the rock-shaft/link parallel lift obvious? → governs claims 12–14.

Second admission, equally damaging. The specification then expressly disclaims a structure closely resembling A10–A17:

"…in a class of planing-machines intended for mill-planing or timber-work… divided upper feed-rolls mounted on a housing have been used, the divided upper feed-rolls being held down by spring-pressure for the independent limited yielding of the divided rolls on the housing, the divided rolls being mounted on a housing, with the housing itself held in place by means of a weighted lever, and so arranged that an offset from the bearings of the divided rolls strikes the housing after the limit of the independent yield of the sectional rolls on the housing has been reached, so as to raise the housing against the action of the weighted lever… and this I do not claim."

Strip the detail: spring cushion + weighted lever + a stop (the housing top) that the roll's bearing part strikes after the spring's independent yield is exhausted, after which the weight's force path changes. That is the same three-element architecture and the same sequential, load-responsive force transmission recited in A13+A15+A16+A17. The applicant has conceded the genus and is claiming a species of it.

---

Step 3 — Level of ordinary skill (Graham step 2)

A person of ordinary skill in the art as of 1899 would be a machine designer in the Cincinnati/Grand Rapids/Massachusetts woodworking-machinery trade, with several years' experience in planer and matcher design, familiar with: the Woodworth lineage (cylinder fixed above an adjustable bed, pressure rollers holding stock to the bed — the founding patent of the industry, 1828, reissued 1845); the standard "wedge-and-incline" bed-elevating arrangement; the use of weighted levers and spring cushions on feed and pressure rolls; compensation/swing gearing for rolls whose journals move; and rock-shafts for equalizing loads at both ends of a roll.

The art was extraordinarily crowded. The patent's own record lists ~25 "similar documents," and the case law of the era (Whitney v. Boston & Albany R., 48 F. 444 (C.C.D. Mass. 1891), concerning US 259,958; the Woodworth/Norcross litigation) shows an active, litigious industry in which every planing-machine refinement was fought over. This supports a high level of ordinary skill — which raises the obviousness bar for the patentee, since a more skilled artisan more readily sees the next step (KSR).

---

Step 4 — Prior-art eligibility (the discipline the record's list lacks)

The "Similar Documents" list is not a prior-art list. Some entries post-date the invention.

Eligible (pre-April 24, 1899, by others):

Not eligible: US 640,833 (Jan. 9, 1900 — and it is Thomas's own even-date companion application), US 642,733 (Thomas's own divisional, filed Aug. 21, 1899 — both "by others"-disqualified and post-dating), US 652,466 (1900, Thomas's own), US 677,243 (1901), US 691,267 (1902), US 701,104 (1902), US 887,021 (1908), US 989,480 (1911), US 1,109,186 (1914). US 185,364 (1876) is also Thomas's own — excluded.

Woodworth (1828/1845) and Hill are eligible as the foundational bed/cylinder/pressure-roll art discussed in the case law.

---

Step 5 — Element mapping

A5/A6/A7 — wedge bed on inclined ways, screw-shifted longitudinally. Met by Wilder (incline planes on the frame mating inverted inclines on the bed underside, shifted by a longitudinal screw to raise/lower the bed for stock thickness, and expressly chosen because it "ensures a solid bearing"). This is the claim's "long wedge giving the table rigidity" almost verbatim. Independently, the applicant admits this element.

A9–A16 — upper roll, per-end bearing, spring interposed between bearing and weighting device, and a stop. Met by US 175,676: bearing O + spring-cushion/buffer O′ + weighted levers D/D′ acting through the buffers + the lever/weight finding "a rest on the top of said housings." Also supported by US 176,918 (spring-loaded yielding feed-roll boxes) and US 180,367 ("controlled by suitable springs or weights").

A17 — the spring-first-then-weight sequencing. Met in substance by the applicant's own admitted prior art (divided rolls held by spring pressure yielding independently up to a limit, then the roll-bearing offset strikes the housing and raises the housing against the weighted lever). This is a two-stage, load-responsive force path in which the spring acts through a first range of motion and the weight takes over in a second range.

Family B — swing housing about the driving shaft, mid-roll gear, parallel lift. The swinging carrier that keeps a moving roll geared is met by RE 6,773 E ("swinging bracket or two-armed hanger H" carrying pinions meshing the roll's internally-toothed pulley) and by US 161,726 (draft rolls on a frame hinged to the main frame). "Parallel lift" of a roll by linkage is met by US 175,676 (opposite-end levers pin-and-slot connected to equalize action) and US 99,008 (yokes/stirrups and weighted levers equalizing pressure at both ends). The "adjustment for the free end of the housing" (set-screws 10) is a routine jacking adjustment; US 176,918 and US 14,304-style screw adjusters for feed-roll boxes are ubiquitous (US 176,918: set-screw m and nut n "by which means the distance between the feed-rolls can be adjusted").

Family C — rock-shaft with pressure-levers at each end, levers pivoted to the frame, links between them, roll bearings on the levers, stops. Met by US 14,038: bearings of the movable roll on joint levers, the levers tied by a connecting rod, weighted bars transferring pressure through the levers, all to move the roll's ends "equally and keep its surface parallel." Substituting a rock-shaft for the connecting rod — so that one shaft acts as a common trunnion for both pressure-levers — is a conventional mechanical equivalent (compare US 180,367's rock-shaft with crank-arms driving two members simultaneously, and US 99,008's lever/stirrup/yoke equalization).

---

Step 6 — The obviousness combinations and the motivation to combine

Combination 1 → Claims 1–7

US 175,676 (Doane & Bugbee) + US 176,918 (Enos) + Wilder, in view of the applicant's admitted prior art.

Why a PHOSITA would combine: Elements A1–A9 are admitted to be the pre-existing class of machine. The only addition is the spring-cushion/weight/stop feeding agency, and US 175,676 — a Cincinnati patent, in the same trade, 23 years earlier — already places a spring-cushion between exactly the two members the claims name (the roll bearing and a weighted lever) and already provides a rest that bounds the lever's travel. US 176,918 supplies the complementary teaching that such springs exist to let the roll "yield as required" while holding the work firmly. Wilder supplies the long-wedge inclined-way bed with a longitudinal screw, and its stated reason — solidity of bearing — is the same reason the applicant gives for his "long wedge giving the table rigidity."

Motivation is not merely generic; it is stated in the art and in the specification itself. The applicant's stated object is a machine that "will give to the stock an especially smooth surface and allow a skimming-cut to be taken," for "furniture, piano, and cabinet work." US 259,958 already frames the same goal ("adapted to regulate the pressure to correspond with the varying thickness of the wood") and was litigated over precisely that commercial value. US 570,867 shows a finishing machine with spring-loaded roll frames arranged to accommodate themselves automatically to varying stock. A designer told "make a finishing planer that skims" and handed a spring-cushion/weighted-lever/rest arrangement would, as a matter of ordinary mechanical reasoning, ask what pressure curve he wants — and the answer (light pressure at shallow cut, increasing pressure as the cut deepens, so the roll is not driven down into the work during a skimming pass) is precisely what a spring in series with a weight naturally produces, because spring force is proportional to deflection while the weight's contribution is constant.

The final increment — relocating the stop so that the weight is normally seated on it rather than pressing continuously, so the weight engages only above a threshold — is a change in the position of a stop and a corresponding change in the point of engagement of a known weight. Under KSR, "a change in the location of a stop" is the paradigm of a predictable, within-skill variation: "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious." The functional clause A17 is a result clause; the result follows from Hooke's law given the structure. No new part is added — the stop already exists in US 175,676.

Combination 2 → Claims 8–11

Combination 1 + US RE 6,773 E + US 161,726 (and US 175,676 for the parallel-lift equalization).

Motivation: Once the in-feed roll is made yielding (as in US 175,676), the immediate engineering problem is that a driven roll cannot simply be allowed to move — its gearing must follow. RE 6,773 E addresses exactly this and teaches the solution: a swinging bracket carrying the intermediate pinions so the drive is maintained through the roll's travel. The remaining step — pivot that swinging carrier on the driving shaft itself so the center distance is geometrically invariant — is the standard, predictable way to keep two gears meshed across a swing; indeed RE 6,773 E is a warning that expansion gearing without proper lever/bracket guidance is "comparatively worthless," i.e., the art was actively working this problem. The applicant's claim that the roll axis lies "in the circle in which it swings" is a tautology of pivoting a housing on the driving-shaft axis — a definitional consequence, not an inventive contribution.

Placing the roll gear at the middle (element B4) with feeding surface to each side is a routine design choice serving the obvious purpose of leaving both ends of the roll shell free for work and keeping the drive in the same plane as the shaft gear — a technique already reflected in US 14,272 (gearing for feed rollers) and RE 6,773 E (internally-toothed pulley at the roll journal).

Note the applicant's own disclaimers: he states he does not claim "broadly, the means for hanging and driving the in-feeding roll," which he says is covered in his separate even-date application (US 640,833). Claims 8–11 nonetheless recite that subject matter in detail — a scoping tension that a validity challenge would exploit.

Combination 3 → Claims 12–14

US 14,038 (Wright, 1856) + US 175,676 or US 99,008, in view of US 180,367.

Motivation: This is the most cleanly met family. US 14,038 already discloses every functional element of the Family C linkage: roll bearings carried on levers, the levers tied by a connecting rod, weights transferred through the levers to the roll, and the express object of moving the roll's ends equally and "keep[ing] their surfaces parallel with the stationary one." The claim's "parallel lift" is the same result, and its stated commercial rationale — that the planed surface "may receive the full impact of the feeding-out roll across its entire width" — is the same uniformity concern US 14,038 addresses. Using a rock-shaft in place of the connecting rod, so that both pressure-levers rock as one on a common journal, is a well-understood mechanical equivalent for tying two levers together (US 180,367 and US 99,008 both use common shafts and yokes for simultaneous, equalized actuation). Substituting a weighted rock-shaft assembly for a weighted connecting-rod assembly, on the same linkage geometry, to achieve the same equalization, is precisely the "combination of familiar elements according to known methods" that KSR holds unpatentable when it "yields predictable results." Claim 14 is even narrower — it merely names the parts — and adds stops, which US 175,676 (the housing-top rest) and US 259,958 (auxiliary support limiting deflection) both show.

Cross-cutting KSR motivations

• Design incentive / market demand: the specification itself identifies a known commercial need (fine furniture, piano, and cabinet work; "the lightest cut with the greatest speed"), a need US 259,958, US 570,867, and US 454,730 ("Machine for forming and polishing door-panels") all show the art was addressing.
• Predictable variation: spring-in-series-with-weight behavior; the engagement point of a stop; the pivot chosen for a driven roll's carrier.
• Known technique / substitution of equivalents: rock-shaft for connecting rod; pivot-on-drive-shaft versus swinging bracket; inclined-plane wedge versus other bed elevators (admitted).
• Obvious to try: given a demand for a two-stage, cut-responsive feed pressure, there were a finite number of predictable ways to arrange a spring, a weight, and a stop — and the art had already used all three of those parts together (US 175,676; applicant's admission).

---

Step 7 — Where the patent is strongest, and how strong the rebuttal is

Intellectual honesty requires identifying the genuine counterarguments.

(a) The one real distinction — the normally-seated weight. In US 175,676 the weighted lever acts continuously through the buffer, and the "rest on the top of the housings" is a downward limit on the lever's free (weighted) end — it stops excess vibration when the roll drops. In US 649,085 the stop supports the weight in a position where the spring alone presses the stock, and the weight is brought into play only when the spring's reaction lifts it off. That is a different load path, and I found no reference I verified that discloses the weight normally seated and inactive. This is the patent's best Hotchkiss/§ 103 argument, and it is a fair one.

Rebuttal: it is a difference in the position of a stop and the engagement point of a known weight, achieved with no new structure, producing a predictable force curve. The applicant's own admitted art already discloses sequential, threshold-based force transmission (spring yields independently to a limit, then the bearing offset strikes the housing and the weight's action takes over). The applicant himself frames the benefit as a matter of degree ("slight pressure… when a skimming cut is being taken and automatically increases the resistance as the cut is increased"), which is the classic language of a change in degree rather than a new principle. Under KSR and pre-1952 "aggregation" doctrine alike, this element alone is unlikely to carry claims 1–7.

(b) Family B is the most defensible in principle, but for a procedural reason rather than a technical one. Claims 8–11 recite a specific packaging (swing housing coaxial with the drive shaft + mid-roll gear + parallel lift + spring/weight/stop + free-end adjustment). A challenger must assemble RE 6,773 E, US 161,726, US 175,676, US 99,008, and US 176,918. Each element is known; the combination is more than a "mere aggregation" because the elements cooperate (keeping a driven, yielding, evenly-loaded roll in a fixed horizontal plane). But the applicant's own disclaimer that the hanging and driving means belongs to a separate application means the examiner may not have intended these claims to carry that weight, and the coaxial pivot is arguably inherent in any swing-housing arrangement.

(c) Parcher teaches the opposite of element A4 — but it doesn't matter. US 180,367 expressly prefers mounting the feed and pressure rolls on the frame rather than the bed, so that raising the bed "causes no increase of strain upon them." That is a genuine teaching away from "lower feed-rolls mounted in the table." However, A4 is admitted prior art in the specification's preamble. An element the applicant concedes is old cannot supply patentability, and a reference's contrary preference cannot resurrect an admitted element. This is a dead end for the patentee.

(d) Claim 13 cannot be assessed. The source omits it entirely. If claim 13 contains a limitation not present in claims 12 and 14 — for example a specific stop geometry or a particular pivot relationship — the analysis could differ. I cannot state an opinion on claim 13 and I am not guessing at its content.

---

Bottom line

Claims 1–7: Likely obvious / invalid. Every element except the two-stage spring-then-weight sequencing (A17) is either expressly admitted prior art or disclosed in US 175,676, US 176,918, US 180,367, US 259,958, and Wilder. A17 is a change in the location of a stop and the engagement point of a known weight, producing a predictable force curve with no new structure — vulnerable under both § 103/KSR and the pre-1952 "change of degree" and "aggregation" doctrines, particularly given the applicant's own detailed admission of a spring-plus-weighted-lever-plus-stop architecture with sequential engagement. Confidence: high.

Claims 8–11: Closer, but still likely obvious. The swing bracket and compensation gearing (RE 6,773 E), the hinged roll-carrying frame (US 161,726), the equalized opposite-end lever loading (US 175,676; US 99,008), and the mid-roll gear (US 14,272; RE 6,773 E) each supply the pieces, and the single design objective — keep a driven, yielding roll in a fixed horizontal plane — supplies a strong, articulated motivation to combine. Confidence: moderate.

Claims 12–14: Most likely obvious. US 14,038 (1856) discloses jointed levers, a connecting rod, weighted bars transferring pressure through the levers, movable roll bearings, and equal, parallel movement of the roll ends — functionally the whole of Family C. Substituting a rock-shaft trunnion for the connecting rod is a recognized equivalent. Confidence: high as to claims 12 and 14; no opinion on claim 13.

The two things that would change my answer: (1) verification of the US RE 9,574 E (Doane, 1881) and US 398,048 / 404,088 (Hutchinson, 1889) contents — the latter two in particular, since a rock-shaft/parallel-lift disclosure there would close Families C entirely; and (2) the actual text of claim 13, which is missing from the source I was given. I hit a tool-call limit before I could retrieve either.

Also flagged: US 640,833, US 642,733, and US 652,466 share this inventor and are not prior art against him, notwithstanding their appearance in the record's similarity list; and US 677,243, US 691,267, US 701,104, US 887,021, US 989,480, and US 1,109,186 all post-date the invention and are ineligible.