US 222895

I'll search for this patent number across the sources you specified.

Let me verify a few specifics — assignee, and the exact filing date (the OCR is garbled).

Search results: US Patent 222,895

Scope note first: I searched both (a) patent/legal databases for US 222,895 and (b) CAFC 2026 dockets. There is no CAFC 2026 docket involving this patent, and that is expected: this patent's term expired December 23, 1896. The only 2026-era hits on the string "222895" are unrelated modern PCT publications (WO 2024/222895, an anti-CD40L antibody case, and WO 2021/222895, an endometrial-cancer diagnostic case). Per your strict rule I am not reconciling these — they are simply different documents that happen to share the numeric string, and neither is US 222,895.

The authoritative record comes from the patent itself (US222895A, Google Patents, fetched 2026-09-30) plus the 19th-century litigation record for the patent.

Bibliographic data

Abstract

There is no abstract. US 222,895 issued in 1879, long before abstracts were required (they arrived with the 1952 Patent Act). The functional equivalent is the specification's statement of invention:

"My invention relates to roller grinding-mills and it consists in a peculiar construction and arrangement of devices for adjusting the rolls vertically as well as horizontally, whereby any unevenness in the wear of the rolls or their journals or bearings may be compensated for, and the grinding or crushing surfaces kept exactly in line."
"The invention also consists in the special devices for separating the rolls when not in action, and in other details."

Problem addressed: roller bearings wear unequally end-to-end and faster on the underside, so rolls lose parallelism and vertical height ("tram").

The seven claims and what they cover

Important framing: this patent carries seven claims, all of which are combination claims standing on their own — none refers back to or incorporates a preceding claim. Under modern nomenclature all seven would be called independent claims, but that label is anachronistic: the statutory independent/dependent distinction did not exist in 1879 practice. Note also that the claims as granted differ from Gray's original claims, because he cancelled and rewrote them after the examiner's August 14, 1879 rejection — a prosecution-history limitation that later proved decisive.

Claim 1 — Gravity fall-back geometry of the movable roll.
The combination of stationary roll B and adjustable roll C mounted in the upper end of pivoted supports D, with the pivots located in advance of the roll's axis, so that the roll falls back away from roll B by its own gravity and stays there when released. (Plain language: the swing-arm pivot is offset forward, so letting go makes the roll drift out of grinding contact on its own — protecting porcelain roll surfaces and preventing grain from wedging while stopped.)

Claim 2 — Swing arm with an eccentric lower pivot plus a reactance at the arm's top.
A roll; an upright swinging arm at each end of that roll; an eccentric, vertically adjustable pivot at the lower end of each arm; and devices acting against the upper end of the arm (i.e., the rod/spring assembly). Caveat: the Google Patents OCR of this claim is garbled — the fetch reads only "the combination of a roll, an upright swinging arm at each end of, and adjusting devices, substantially such as shown, to regulate the tension of the spring." The fuller wording above is reconstructed from the parallel recitation of the same claim in Consolidated Roller-Mill Co. v. Coombs, 39 F. 25 (E.D. Mich. 1889) and the Supreme Court's opinion; treat the middle limitation as moderately confident, not verbatim.

Claim 3 — Vertically adjustable pivot that doubles as the roll's support.
The combination of a roll and upright swinging arms whose lower ends are mounted on vertically adjustable pivots, those pivots "thus serving both to sustain and adjust the rolls." (Plain language: one part — the eccentric/block/screw pivot — both carries the arm's weight and sets the roll's vertical height, i.e., the tram adjustment.)

Claim 4 — The rod G / spring sub-combination (the commercially important claim).
In combination with the movable roller bearing: the rod G; an adjustable stop device to limit the bearing's inward movement; an outside spring urging the bearing inward; and adjusting devices to regulate the spring's tension. (Plain language: a horizontal tie rod running from the fixed bearing to the top of the swing arm, with a spring outside the bearing that pushes the roll toward its mate and can be tensioned by a nut — and which yields if a nail or stone passes between the rolls.)

Claim 5 — The adjusting rod with stops at both ends.
In combination with the roller bearing: the adjusting rod provided at one end with a stop limiting inward movement, a spring, and means for adjusting the spring; and at the other end a stop and holding devices. (Plain language: the same rod, but claimed end-to-end — spring-and-tension-adjuster on one side, a shoulder stop plus nut on the other side, so the roll can be released to fall back and then returned to exactly its prior position without re-adjusting.)

Claim 6 — The specific hardware combination.
The combination of bearing D, rod G, nut l, spring H, nut j, stop n, and nut o — a "picture claim" reciting the exact parts by letter. Caveat: the lettering is OCR-corrupted in the fetched text ("nut 1", "stop a", "nut 0"); the Supreme Court opinion renders it as "bearing D, rod G, nut l, spring H, nut j, stop n, and nut O." Treat the identifiers as letter-label, not digit, predictions.

Claim 7 — The eccentric sleeve pivot with its locking screw.
The combination of the roller, the upright supporting arm D, the eccentric sleeve F on which the arm's lower end is mounted, and the screw E applied to secure the sleeve against rotary movement. (Plain language: the wrench-headed eccentric sleeve that raises/lowers the arm, clamped by screw E against the frame so it becomes the pivot for horizontal adjustment.)

Litigation / validity context you should not omit

This is not a clean patent. Its scope was materially narrowed during prosecution (examiner cited English patent No. 3,328 of 1877 — the "Lake" patent, covering Nemelka of Simmering, Austria — and Gray cancelled all original claims, adding "peculiar" and "special" to his statement of invention). The resulting narrow construction defeated infringement in Consolidated Rolling-Mill Co. v. Barnard & Leas Mfg. Co., 156 U.S. 261 (1895) (Supreme Court, not CAFC), where the Court held the defendant's upright rod with an inside spiral spring was not the equivalent of Gray's horizontal rod G, and found no infringement of claims 4, 5, and 6; the companion Gray patent 238,677 was held void for want of novelty. Earlier, Consolidated Roller-Mill Co. v. Coombs, 39 F. 25 (E.D. Mich. 1889), and the circuit decision at 43 F. 527 (N.D. Ill.), reached similar non-infringement results. The Court expressly characterized Gray's patent as not a pioneer patent and therefore not entitled to liberal equivalents.

Confidence statement

• High confidence: patent number, inventor, title, issue date, expiration date, the seven claim texts' substance, the rod-G/spring architecture, and the Supreme Court outcome.
• Moderate confidence / flagged uncertainty: the exact filing day within July 1879 (OCR shows "July so, 1879"; I infer July 30); the verbatim wording of claim 2 (OCR loss); the letter labels in claim 6 (OCR substituted digits); and the existence of a recorded assignment to Consolidated Roller-Mill Company (inferred from litigation, not confirmed from an assignment record).
• Negative finding: no CAFC docket (2026 or otherwise) involves US 222,895.

I'll search for litigation involving US patent 222895 specifically.

I found a lead. Let me dig deeper into the specifics of this 1895 Supreme Court case and any related lower court proceedings.

I found a reference to a second case. Let me investigate that and look for the lower court docket details.

Litigation involving U.S. Patent No. 222,895

Important scoping note: U.S. Patent 222,895 ("Improvement in roller grinding-mills," issued Dec. 23, 1879 to William D. Gray; application filed July 30, 1879) is an 1879 patent that expired around Dec. 23, 1896. All known litigation is therefore 19th‑century, pre‑PACER, and predates modern electronic docketing. I found no modern litigation. My searches also returned several unrelated patents with similar-looking numbers — e.g., U.S. Patent 4,222,895 (Hockerson shoe patent, Hockerson-Halberstadt v. Nike/Converse), the 10,221,780 family asserted by Champion Power Equipment, and WIPO publications WO2021/222895 and WO2024/222895. Those are not U.S. Patent 222,895 and I have excluded them.

The patent was asserted in a series of infringement suits brought by the Consolidated Roller-Mill Company (and one by another party). The confirmed matters are below.

1. Consolidated Roller-Mill Co. v. Coombs

• Plaintiff: Consolidated Roller-Mill Co.
• Defendant: Coombs
• Jurisdiction: U.S. Circuit Court, E.D. Michigan (Judge Brown)
• Case number: Not available (pre‑PACER); reported at 39 F. 25
• Filing date: Not confirmed from available sources (opinion circa 1888–1889)
• Outcome: Plaintiff won on liability. The court upheld the validity of claims 2, 3, 4, and 5 of the Gray patent (rejecting the Lake/Nemelka anticipation defense) and held the defendant's "Mawhood" machine infringed.
• Source: https://law.resource.org/pub/us/case/reporter/F/0040/0040.f.0305.pdf (discussing Roller-Mill Co. v. Coombs, 39 Fed. Rep. 25)

2. Consolidated Roller-Mill Co. v. George T. Smith Middlings Purifier Co.

• Plaintiff: Consolidated Roller-Mill Co.
• Defendant: George T. Smith Middlings Purifier Co.
• Jurisdiction: U.S. Circuit Court, E.D. Michigan (Judge Brown)
• Case number: Not available; reported at 40 F. 305
• Filing date: Decided Oct. 7, 1889 (motion for preliminary injunction)
• Outcome: Defendant prevailed on this motion — preliminary injunction denied. The court held the Gray patent "does not cover a machine having no lever or swinging arm which may be raised or lowered for a vertical adjustment, and no rod corresponding to the rod G . . . fitted at either end with devices for horizontal adjustments." The court also addressed a res adjudicata/comity objection arising from the Wisconsin case (below).
• Source: https://law.resource.org/pub/us/case/reporter/F/0040/0040.f.0305.pdf

3. Allis v. Freeman

• Plaintiff: Allis
• Defendant: Freeman
• Jurisdiction: U.S. Circuit Court, W.D. Wisconsin
• Case number: Not available
• Filing date: Not confirmed (decree circa 1886–87; appeal docketed in the Supreme Court in October 1887)
• Outcome: Bill dismissed — the court appears to have been of the view that the Gray patent was anticipated by the English patent to Lake (Nemelka). An appeal was perfected and docketed in the U.S. Supreme Court in Oct. 1887 but was later dismissed on the appellant's motion/consent. The Gray patent (No. 222,895) was one of several patents at issue.
• Source: https://law.resource.org/pub/us/case/reporter/F/0040/0040.f.0305.pdf

4. Consolidated Roller-Mill Co. v. Barnard & Leas Manuf'g Co. (Circuit Court)

• Plaintiff: Consolidated Roller-Mill Co.
• Defendant: Barnard & Leas Manufacturing Co.
• Jurisdiction: U.S. Circuit Court, N.D. Illinois (Judge Blodgett)
• Case number: Not available; reported at 43 F. 527
• Filing date: Not confirmed (opinion 1890; rehearing overruled July 14, 1890)
• Patents asserted: Gray No. 222,895; Gray No. 238,677; Odell reissue No. 10,139; Birkholz No. 269,623
• Outcome: Defendant prevailed. The court held the Gray patent No. 222,895 was not infringed, Gray No. 238,677 was void for want of novelty, the Odell reissue was void for want of invention, and the Birkholz patent was not infringed. Rehearing overruled July 14, 1890.
• Sources: https://law.resource.org/pub/us/case/reporter/F/0043/0043.f.0527.1.pdf ; https://www.courtlistener.com/opinion/[9310612](/patent/9310612)/consolidated-roller-mill-co-v-barnard-leas-manufg-co/

5. The Roller Mill Patent, 156 U.S. 261 (1895) — U.S. Supreme Court

• Caption below: Consolidated Roller-Mill Co. v. Barnard & Leas Manuf'g Co.
• Plaintiff/appellant: Consolidated Roller-Mill Co.
• Defendant/appellee: Barnard & Leas Manufacturing Co.
• Jurisdiction: U.S. Supreme Court (on appeal from the U.S. Circuit Court for the Northern District of Illinois)
• Case number: No. 70 (Supreme Court term docket); reported at 156 U.S. 261
• Dates: Argued Nov. 12, 1894; decided Feb. 4, 1895
• Outcome: Affirmed for the defendant. The Court held that the invention protected by letters patent No. 222,895 "is not infringed by the machine used by the defendant in error," and that letters patent No. 238,677 "are void for want of novelty." The Court agreed the Nemelka Austrian and French patents of 1875 and the Nemelka/Lake English patent of 1877 anticipated the Gray adjustments, and that prosecution-history limitations narrowed the claims.
• Sources: https://supreme.justia.com/cases/federal/us/156/261/ ; https://tile.loc.gov/storage-services/service/ll/usrep/usrep156/usrep156261/usrep156261.pdf ; https://www.courtlistener.com/opinion/94092/the-roller-mill-patent/

Summary

U.S. Patent 222,895 (Gray) was litigated in at least five reported proceedings in the 1880s–1890s, all in the federal circuit courts and once in the Supreme Court. The patent owner's fortunes were mixed: it won in Coombs (patent upheld and infringed) but lost in George T. Smith Middlings Purifier (no infringement, PI denied), in Barnard & Leas before Judge Blodgett in N.D. Illinois, and on appeal in the Supreme Court in The Roller Mill Patent, 156 U.S. 261 (1895), which held the machine did not infringe. The companion Gray patent 238,677 was held void for want of novelty. The Allis v. Freeman suit in W.D. Wisconsin was dismissed against the patent holder.

Confidence and limitations

• High confidence: the parties, jurisdictions, citations (39 F. 25; 40 F. 305; 43 F. 527; 156 U.S. 261; Supreme Court No. 70), and outcomes above, which are drawn from the published opinions.
• Cannot confirm: modern-style case numbers (the federal courts did not use them in the modern format; the N.D. Ill. Circuit Court dockets were not located in the sources searched), and exact filing dates for the lower-court suits — these are estimated ranges only. PACER/CAFC databases do not cover 1880s–1890s circuit court filings, and no post‑1900 litigation (including any IPR, reissue, or later infringement action) involving U.S. 222,895 was found.
• The Google Patents page for US222895 shows a "family litigation" link (Darts‑IP family ID 2292290), but no case details were retrievable, consistent with the historical nature of the disputes described above.

Proceedings overview

Total AIA trial proceedings on US 222895: 0. As of the most recent USPTO Open Data Portal ingest, there is not a single IPR, PGR, or CBM petition on file against this patent. The breakdown by status is: active — 0; claims invalidated — 0; claims sustained — 0; settled — 0; institution denied — 0.

The bottom-line defensive posture is not "hardened patent" and not "claims canceled" — it is "there is nothing here to defend against." US 222895 issued 1879-12-23, ran its 17-year term, and expired 1896-12-23 (Google Patents legal status: "Expired - Lifetime," anticipated expiration 1896-12-23). No AIA trial is possible as a practical matter: PGR is unavailable because the effective filing date (1879-07-30) predates 2013-03-16 by 134 years; CBM is unavailable both on subject-matter grounds (a mechanical grain roller-mill is not a covered business method patent) and because the CBM program sunset on 2020-09-16. IPR is theoretically available against any patent regardless of filing date, but there is no live owner, no enforceable term, and no damages window (§ 286 reaches back only six years from a complaint). If a demand letter cites US 222895 today, the correct response is not an IPR petition — it is a one-paragraph letter explaining that the patent expired in the nineteenth century.

There are, therefore, no proceedings to enumerate in the required per-proceeding format. Rather than fabricate numbers (I will not), the section below covers the only adjudicative history this patent has — 19th-century federal court litigation — clearly labeled as not AIA trials.

⚠️ Not PTAB proceedings — the actual adjudicative record on US 222895

The proceedings below are Article III / circuit court actions under the Patent Act of 1870, not AIA trials. They are included because they are the only claim-level outcomes that exist for this patent, and because they may be what a search hits when someone looks for "litigation" on this number.

Consolidated Roller-Mill Co. v. Coombs, 39 F. 25 (C.C.E.D. Mich. 1889) — https://law.resource.org/pub/us/case/reporter/F/0039/0039.f.0025.pdf

• Holding: Patent 222,895 valid and infringed (the "Mawhood Machine"), claims 2, 3, 4, and 5 sustained. The court declined to read the patentee's Patent Office amendments as a surrender of the doctrine of equivalents.
• Defensive value today: None. This is the outlier pro-patent outcome, and it was a co-ordinate-circuit comity problem for the courts that followed.

Consolidated Roller-Mill Co. v. George T. Smith Middlings Purifier Co., 40 F. 305 (C.C.E.D. Mich. 1889) — https://law.resource.org/pub/us/case/reporter/F/0040/0040.f.0305.pdf

• Holding: Preliminary injunction denied; patent "does not cover a machine having no lever or swinging arm which may be raised or lowered for a vertical adjustment, and no rod corresponding to the rod G… fitted at either end with devices for horizontal adjustments."
• Also held that a prior invalidity decree is not in rem and does not estop a later suit against a different defendant (distinguishing Allis v. Freeman, C.C.W.D. Wis., no opinion filed).

Consolidated Roller-Mill Co. v. Barnard & Leas Mfg. Co., 43 F. 527 (C.C.N.D. Ill. 1890) — https://law.resource.org/pub/us/case/reporter/F/0043/0043.f.0527.1.pdf

• Holding: Bill dismissed. The printed headnote says the Gray patents "are anticipated by the Nemelka Austrian and French patents of 1875, and the Nemelka Lake English patent of 1877"; Judge Blodgett went further off the bench, writing that "Mr. Gray in effect invented nothing." Rehearing denied 1890-07-14. Note the headnote's "anticipated" language is broader than what the Supreme Court later treated as the actual disposition — see below.

The Roller Mill Patent, 156 U.S. 261 (1895) (No. 70; argued 1894-11-12; decided 1895-02-04) — https://www.courtlistener.com/opinion/94092/the-roller-mill-patent/ | https://supreme.justia.com/cases/federal/us/156/261/

• Holding: "The invention protected by letters patent No. 222,895… is not infringed by the machine used by the defendant in error." The Court construed Gray's claims narrowly against the prior art and the Patent Office prosecution history (the 1879-08-14 rejection over English patent No. 3,328 of 1877 to Lake, which forced Gray to withdraw all original claims and rewrite them as specific combinations), held the patent "is not a pioneer patent and is not entitled to that liberality of construction," and affirmed dismissal on non-infringement — not on invalidity.
• Companion patent: Patent No. 238,677 (Gray, 1881-03-08) was held void for want of novelty in the same appeal.
• Claims addressed: The Court expressly analyzed claims 4, 5, and 6 (the rod-G combination claims) because those were the only ones plaintiff asserted. It described the patent as containing seven claims, with claims 2–3 directed to vertical/horizontal adjustment and claims 4–5 to the rod-G devices.
• Claims NOT decided: The Court did not invalidate any claim of 222,895; it never reached validity. Claims 1, 2, 3, and 7 were not adjudicated.

Google Patents "litigation" flag: The US222895A page carries a Darts-ip link for family ID 2292290 under "Family has litigation." I could not independently verify what is in that dataset; the only litigation I can confirm is the 1889–1895 federal record above. Treat that flag as a pointer to 19th-century infringement suits, not to PTAB activity.

Strategic summary

Claim status — nothing is "canceled," because no tribunal ever canceled anything, and no claim is enforceable either. US 222895 has seven claims as printed. The Google Patents text layer shows only claims 1, 2, 5, 6, and 7 — claims 3 and 4 are missing from that OCR rendering, so anyone working from the Google Patents page is working from an incomplete claim set. The 1895 Supreme Court decision confirms the patent contains seven claims and quotes claims 4, 5, and 6. On the merits: no claim of 222895 has ever been held invalid in a published opinion; no claim has been held infringed except in Coombs (39 F. 25) as to the Mawhood machine. The 1895 affirmance was on non-infringement grounds only, and the Supreme Court deliberately declined to reach the Nemelka anticipation question. So if you see a characterization that "claims 4–6 were invalidated in 1895," that is wrong — the Supreme Court affirmed a dismissal for non-infringement.

Estoppel landscape. Section 315(e)(2) estoppel is irrelevant here — there is no IPR to generate it, and no petitioner to bind. The relevant bars are temporal and remedial: 35 U.S.C. § 286 limits recovery to six years before a complaint, and the patent has been expired since 1896-12-23. Practically, every prior-art ground ever assembled against this patent — the Nemelka Austrian patents (1875-01-15, 1875-05-22), the Nemelka French patent (1875-06-23), and the English Lake patent No. 3,328 of 1877 — remains available to anyone, unencumbered. It is worth noting for completeness that if (contrary to all indications) an IPR were ever filed, § 311(b) limits IPR grounds to § 102/§ 103 on patents and printed publications only; the 1879-08-14 and 1879-08-19 Patent Office rejections documented in the file wrapper would be classic § 325(d) discretionary-denial material.

Pattern signals. No petitioner has filed anything, because there is no one to petition against and nothing to win. There is no defensive aggregator (Unified Patents or similar) anywhere in the chain — there is no chain; the patent term ended under the pre-1930 17-year term and no assignment activity is on record. The patent owner of record (William D. Gray, of Milwaukee, Wisconsin) died long ago, the last-named commercial plaintiff (Consolidated Roller-Mill Company) has been defunct for over a century, and no maintenance-fee, revival, or assignment trail exists. The absence of PTAB activity is not a signal of a hardened patent here — it is a signal that the paper is a museum piece.

Recommended next steps

1. If you have received a demand citing US 222895, do not file an IPR. IPR is available in principle against expired patents, but it would cost six figures to obtain a decision that the patent expired in 1896. The efficient response is a letter citing US 222895's own face: issuance 1879-12-23, expiration 1896-12-23 ("Expired - Lifetime" per https://patents.google.com/patent/US222895/en), plus 35 U.S.C. § 286's six-year damages bar.
2. If you must litigate, your strongest authority is the 1895 Supreme Court decision, which construed the claims narrowly and held no infringement: The Roller Mill Patent, 156 U.S. 261, 271 (1895) — "we think the circuit court was correct in holding that defendant's machine was not an infringement of the gray patent. Should this device be adjudged an infringement we should not know where to draw the line." Full text at https://www.courtlistener.com/opinion/94092/the-roller-mill-patent/. Note that the Court only reached claims 4, 5, and 6; claims 1, 2, 3, and 7 have no controlling construction.
3. Do not cite claims 3 and 4 from the Google Patents page — they are absent from that rendering. Pull the printed patent from the USPTO PatentCenter link at https://patents.google.com/patent/US222895/en or the 156 U.S. 261 appendix before quoting any claim verbatim.
4. No trial-stage milestones to track. There is no institution decision deadline, no oral hearing date, and no statutory one-year FWD clock, because there is no proceeding. PTAB E2E (https://ptab.uspto.gov) and the USPTO PTAB Decisions database (https://developer.uspto.gov/ptab-web) can be checked to confirm the null result at any time.
5. Verify before relying on the negative. I searched for PTAB activity and found none; the ODP structured data likewise returns none. I state this with high confidence. I could not verify the contents of the Darts-ip family ID 2292290 dataset referenced on the Google Patents page, so I cannot rule out that it contains a records entry I have not seen — but a Darts-ip "family litigation" flag is a commercial dataset pointer, not an AIA trial, and no PTAB proceeding number exists in the public record for US 222895.

I'll search the USPTO assignment records and related sources for this patent's ownership chain.

Let me try the USPTO assignment search endpoints and verify their historical coverage.

The critical coverage fact is confirmed. Now let me establish what the historical ownership record shows outside the August-1980-onward database.

Scoping note up front

This is a pre‑1980 patent with no retrievable electronic assignment record. USPTO's Patent Assignment Search / Assignment Center (assignmentcenter.uspto.gov, mirrored at assignment.uspto.gov) contains assignments recorded from August 1980 forward only; MPEP 301.01 provides that the pre‑1980 "historical database of assignments is maintained separate and apart from the database of assignments affecting title recorded since 1980," and that older records are not exposed to the online search-by-patent-number interface. US 222,895 issued 1879‑12‑23, was filed 1879‑07‑30, and its term expired 1896‑12‑23 — roughly 84 years before the searchable window opens. Any 19th‑century assignment is in the microfilm historical file or in the litigation record, not in Assignment Center.

That means no reel/frame data exists for me to report, and I have not invented any. What follows distinguishes recorded assignments (none) from documented ownership evidence (court opinions, which are not assignment records).

Inventors

• Sole named inventor. The patent is granted to "WILLIAM D. GRAY, OF MILWAUKEE, WISCONSIN" personally, with no assignee named on the face — normal for 1879, when an assignee was reflected only if the assignment was recorded and the patent printed accordingly.
• No co‑inventors. The two names appearing under the signature block — "WILLIAM McLEAN, Jr." and "O. M. FRENCH" — are witnesses, not inventors. Do not mis-key them into an assignor table.
• Gray was a serial filer in the same art: he also obtained US 238,677 (issued 1881‑03‑08) on a related roller‑mill spreading/adjusting mechanism. Two filings in the same narrow mechanical space within ~18 months indicates an active designer, not a one-off filer.
• Unusual-pattern check (inventor departures preceding a fire sale): not assessable. There is no employment, no assignment-from-inventor record, and no 12‑month window to measure — the entire signal requires 1980+ data that does not exist here. I cannot call this present or absent.

Original assignee

None. The patent issued to the inventor personally; it was inventor-owned at grant.

The patent's front matter (and the Google Patents record) shows no assignee field. Google Patents' legal-events block for US222895 likewise shows only granted (1879‑12‑23) and anticipated expiration (1896‑12‑23) — no assignment events. That absence is real, not a gap in my search.

What the litigation record shows about downstream ownership (secondary evidence, not recorded assignments):

• The patent was owned by "Allis" by the mid‑1880s. Consolidated Roller-Mill Co. v. Richmond City Mill‑Works states: "Prior to the commencement of the Michigan suit, Allis, the then owner of the Gray patent, sued Freeman for its infringement in the western district of Wisconsin." (CourtListener: https://www.courtlistener.com/opinion/[8853526](/patent/8853526)/consolidated-roller-mill-co-v-richmond-city-mill-works/, opinion paginated at *475.) Given the Milwaukee venue and the machinery trade, this is most likely Edward P. Allis / E. P. Allis & Co. of Milwaukee — the firm that became part of Allis‑Chalmers in 1901 — but I flag that identification as probable, not confirmed.
• After losing at the circuit level in Wisconsin, Allis "assigned the patent to the parties, or some of them, who became complainants in the Michigan suit, and they dismissed the appeal" (same opinion). That transferee group is the Consolidated Roller‑Mill Company, plaintiff in Coombs, George T. Smith Middlings Purifier, Richmond City Mill‑Works, and Barnard & Leas.
• Consolidated Roller‑Mill Co. did not manufacture. In Coombs, the defendant's fourth defense was "that the plaintiff has never been engaged in the manufacture, sale, or use of the alleged inventions covered by its patents." In Richmond City Mill‑Works the court observed: "The defendant is a manufacturer, and the complainants are not." It asserted a pool of patents by different inventors — Gray 222,895, Gray 238,677, Odell reissue 10,139, Birkholz 269,623, Dowling 289,518, Marmon 274,508 — i.e., a licensing/assertion vehicle, 1880s style.
• Current status: none applicable. The patent entered the public domain on 1896‑12‑23. There is no current owner to check against SEC filings, RPX, or Unified Patents directories. (SEC postdates this patent by 55 years; there is no 10‑K/8‑K analogue.)

Assignment timeline

The USPTO Assignment Center has no records for US 222,895. Searching by patent number returns nothing, for the August‑1980 coverage reason above. Per the task rules, that is itself the finding: no recorded post-issuance assignments are retrievable for this patent, and I will not fabricate reel/frame entries.

Below is the unrecorded ownership history as documented in the published opinions. Every entry is labelled as such and has no reel/frame.

• Undated, between 1879‑12‑23 and c. 1886 (executed) / not recorded — Reel/frame: none

  • Conveyance: Assignment (unrecorded; inferred from opinions)
  • Assignor: William D. Gray (individually)
  • Assignee: "Allis" — probably Edward P. Allis, Milwaukee
  • Correspondent: not applicable — no recordation, therefore no correspondent of record
  • Context: sale/transfer of the patent out of the inventor's hands to a Milwaukee machinery manufacturer

• c. 1887–1888 (executed) / not recorded — Reel/frame: none

  • Conveyance: Assignment (unrecorded)
  • Assignor: Allis
  • Assignee: Consolidated Roller‑Mill Company and/or unnamed co‑complainants
  • Correspondent: none of record. For clarity: Rodney Mason (plaintiff's counsel in Coombs) and Joseph G. Parkinson / Robert H. Parkinson (defendant's counsel) are litigation counsel, not recordation correspondents — do not conflate them with the assignment-correspondent signal.
  • Context: transfer-to-asserter, executed while the Wisconsin appeal was pending; the new owners then dismissed the appeal and brought the Michigan suits (source: Richmond City Mill‑Works, *475)

• 1889‑05‑20 (decision date) / no recording — Reel/frame: none

  • Conveyance: none — litigation event only
  • Consolidated Roller‑Mill Co. v. Coombs, 39 F. 25 (Gray patent claims 2–5 upheld; Mawhood machine infringing)
  • Context: assertion by the non-manufacturing owner against a maker

• 1890 / no recording — Reel/frame: none

  • Conveyance: none — litigation event only
  • Consolidated Roller‑Mill Co. v. Barnard & Leas Mfg. Co., 43 F. 527 (N.D. Ill., Blodgett, J.): not infringed; companion Gray 238,677 void for want of novelty

• 1895‑02‑04 / no recording — Reel/frame: none

  • Conveyance: none — judicial disposition
  • The Roller Mill Patent, 156 U.S. 261 (No. 70): affirmed, no infringement; 238,677 void

• 1896‑12‑23 / statutory event — Reel/frame: none

  • Conveyance: none — term expiration, chain ends

Timeline diagram

timeline
    title Ownership and assertion of US 222895
    1879 : Gray files in July
         : Patent 222895 issues Dec 23
         : No assignee named on the patent face
    1881 : Gray obtains related patent 238677
    1886 : Allis is owner of the Gray patent
         : Allis loses in Wisconsin against Freeman
    1887 : Wisconsin appeal docketed then dropped
    1888 : Allis assigns patent to Michigan plaintiffs
         : Consolidated Roller Mill Co begins assertion
    1889 : Owner wins Coombs suit
         : Owner loses George T Smith motion
    1890 : Barnard and Leas wins at circuit
    1895 : Supreme Court finds no infringement
    1896 : Patent term expires

Two events in 1879, 1886 and 1888 are stacked with leading colons, as required. Mid‑1880s year labels are approximate — the underlying documents give no date for the Gray→Allis transfer.

NPE / troll-pattern signals

Because the recorded chain is empty, most signals are assessed as "cannot be evidenced," not "proven absent." I distinguish the two.

1. Shell-entity transfer — Not present (as a recorded fact). No 1980+ records exist, so no LLC-of-record evidence is possible. Historical analogue worth noting, cleanly separated: the 1887–88 transfer ran from an owner to Consolidated Roller‑Mill Co., which courts expressly found was not a manufacturer ("the complainants are not"; Richmond City Mill‑Works, *475). That is the 19th‑century functional equivalent of an operating-to-assertion transfer — but there is no recorded document, entity name suffix, or registered-agent address to cite, so it does not satisfy the signal's evidentiary standard.

2. Known asserter in the chain — Not present. No assignee in the documented chain appears on any modern NPE directory (Acacia, Marathon, IV, IPNav, Wi‑LAN, Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, DGC, Spangenberg entities, or Unified/RPX high-frequency plaintiff lists). Those lists are contemporary; they cannot reach 1880s entities, and I will not force a match.

3. Repeat correspondent across the chain — Not present / not assessable. Zero recorded assignments means zero correspondents of record to compare. I can state affirmatively that no attorney name appears as a recordation correspondent anywhere in this chain. Litigation counsel names are not a substitute.

4. Cascading transfers — Unclear (historically suggestive, evidentially thin). The opinions support at least two transfers inside roughly two years (Gray→Allis sometime by the mid‑1880s; Allis→Michigan complainants c. 1887–88), and the second immediately preceded concentrated assertion activity 1888–1890. But: no execution dates, no recording dates, no shared-correspondent or shared-address evidence. No reel/frame — so I cannot convert this into a finding.

5. Pre-litigation transfer — Unclear, leaning structurally present. The Allis→Michigan-complainant assignment is described as occurring after the Wisconsin loss and before the Michigan suits, and its evident purpose was to consolidate standing in asserting hands so the appeal could be abandoned and fresh suits filed. That is the mechanics of a pre-litigation transfer. However, ABSENT exact execution and suit-filing dates I cannot test the "<6 months" threshold, and no assignment document is cited. Marked unclear rather than present.

6. Bankruptcy fire-sale — Not present. No evidence of any insolvency proceeding involving Gray, Allis, or Consolidated Roller‑Mill Co. for which this patent was sold. Note that the modern Chapter 7/11 framework (1898 Bankruptcy Act) postdates the relevant window, so the modern signal has no vehicle here. Consolidated Roller‑Mill Co.'s ultimate corporate fate is not determinable from the sources I searched.

7. Privateering — Unclear. The pattern (an entity holding a pool of patents by Gray, Odell, Birkholz, Dowling, and Marmon and suing mill manufacturers) is structurally privateering-like, but the defining element — an operating company transferring patents to an NPE that asserts on its behalf, surfaced via SEC filings or EFF/Patent Progress coverage — cannot be shown. SEC did not exist; no such coverage exists for 1889. Marked unclear, not present.

8. Defensive aggregator (anti-NPE) — Not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN — none of which existed. The patent was neutralized instead by judicial decision (Barnard & Leas, 43 F. 527, 1890; affirmed 156 U.S. 261, 1895, holding non-infringement and voiding companion Gray 238,677) and then by expiration on 1896‑12‑23. This is a distinct termination mode from the defensive-aggregator signal.

Verdict

Insufficient data.

Justification: USPTO's assignment database covers only records from August 1980 onward (MPEP 301.01), while this patent issued 1879‑12‑23 and expired 1896‑12‑23 — so there are zero reel/frame entries to evaluate and no correspondent-of-record, correspondingly defeating signals 1, 2, 3, and 8 on the available evidence. The only ownership evidence is unrecorded, non-reel/frame narrative in 19th‑century opinions — an "Allis" ownership followed by an 1887‑88 assignment to the non-manufacturing Consolidated Roller‑Mill Company, which then asserted against manufacturers ("the complainants are not" manufacturers; Richmond City Mill‑Works, *475) — a structurally troll-like pattern that nonetheless cannot be graded as an NPE signal without recorded assignments or dates.

A closing cross-check against the earlier litigation section: the Richmond City Mill‑Works opinion independently corroborates that section's account of the Allis v. Freeman appeal being dismissed on the appellant's side — and adds the missing causal link (the appeal was dropped because the patent had been assigned away to the Michigan complainants). That opinion was not in the earlier list of five proceedings, so the litigation set arguably expands to six matters; I flag it as an addition rather than a contradiction.

Verification links:

• USPTO Assignment Center: https://assignmentcenter.uspto.gov/
• Legacy search interface: https://assignment.uspto.gov/patent/index.html (search by patent number → no results expected)
• MPEP 301.01 basis for the 1980 coverage cutoff (secondary discussion): https://blueironip.com/ufaqs/how-are-historical-patent-assignment-records-maintained/
• Google Patents legal events for US222895 (grant + expiration only, no assignments): https://patents.google.com/patent/US222895/en
• Ownership evidence: https://www.courtlistener.com/opinion/8853526/consolidated-roller-mill-co-v-richmond-city-mill-works/
• Disposition: https://supreme.justia.com/cases/federal/us/156/261/

One caution up front: anyone querying Assignment Center for "222895" will also surface unrelated modern records (e.g., US 4,222,895; the 10,221,780 family; WO2021/222895; WO2024/222895). Those are not this patent and must be excluded from any chain-of-title table.

I'll search for the patent's citation records and prior art.

Now let me pin down the specific prior-art references named in the record, and check whether the patent's own citation record lists any U.S. patents.

Let me verify the Nemelka U.S. patent and check for any other prior-art patents named in the record, plus the Google Patents citation data.

Prior Art of U.S. Patent 222,895 — Citation Analysis

0. Two framing points before the table (read these first)

(a) There is no machine-readable citation record for this patent, and I could not verify one exists at USPTO. US 222,895 issued 23 December 1879. Patents of that era were printed with no "References Cited" section, no examiner's citation list, and no applicant IDS. The USPTO's Patent Public Search / Patent Center does not carry prosecution or citation metadata for pre-1980 (indeed pre-1930s) cases in retrievable form; only the patent image/specification is available. Consistent with that, the Google Patents record for US222895A (fetched 2026-09-30) contains no "References Cited" field, no "Cited By" field, and no "Patent Citations" table — only a "Similar Documents" list and a prior-art-keyword list (roll; mills; rolls; nut; roller). I therefore could not "look at each patent citation for 222895" in the modern sense, because none is recorded. What follows reconstructs the citation set from the two places where it actually exists: (i) the prosecution rejection of record, and (ii) the prior-art record made in the 1889–1895 infringement litigation.

(b) Statutory anachronism. 35 U.S.C. § 102 did not exist in 1879. The governing law was R.S. §§ 4886/4920. The § 102 labels below are used as the task requests, but they are applied by analogy only. I use two critical dates: the application date — "made in July, 1879" (per The Roller Mill Patent, 156 U.S. 261, 265; Google's printed line reads "application filed July so, 1879," almost certainly July 30, flagged as OCR-uncertain) and the issue date, 23 December 1879. Note that Google's displayed "Prior art date 1879-12-23" is the issue date, not a filing date.

1. The only citation actually of record in prosecution

Reference A — English patent to Lake, No. 3,328 of 1877 ("the Lake English Patent")

• Full citation: English patent No. 3,328 of 1877, granted to Lake (as grantee/agent) for the invention of Lorenz Nemelka of Simmering, Austria; complete specification dated/accepted 28 February 1878. As the Court describes it: "a patent in England, issued to Lake, February 28, 1878." The "of 1877" numbering is the British practice of numbering by year of application.
• Publication/filing date: applied for 1877; dated 28 February 1878. Both pre-date Gray's July 1879 filing by more than a year.
• Brief description: Multiple forms (Figs. 11–15 among others) of a roller mill having both a vertical adjustment (a roll adjustable by a set screw underneath it) and a horizontal/parallelism adjustment (a sliding bracket carrying the bearing of a shaft that works in an eccentric journal, operated by a lever pivoted on the shaft, by which the rolls are opened when not in operation). Some forms show a yielding capability via an india-rubber buffer at the lower end of a long descending arm of the movable bearing.
• § 102 mapping: This is the key reference and the only one the examiner cited. It was the express basis of the 14 August 1879 rejection: the invention "is not generic in view of the English patent No. 3,328, of 1877, this being known as the 'Lake English Patent.'" It was treated as anticipating the original claims 1–5 (broad vertical + horizontal adjustment), which Gray then withdrew in their entirety. Under § 102(a)/(b) it is the strongest single anticipation candidate against the broad concept, and — because it is more than one year before the July 1879 filing — an outright § 102(b) statutory bar as to that concept.
• Caveat: Gray submitted to the rejection rather than contesting it, so there is no adjudicated finding that Lake alone anticipated any granted claim.

Sources: https://supreme.justia.com/cases/federal/us/156/261/ ; https://tile.loc.gov/storage-services/service/ll/usrep/usrep156/usrep156261/usrep156261.pdf ; https://www.courtlistener.com/opinion/94092/the-roller-mill-patent/

2. The Nemelka family — the substantial prior art at trial

The Patent Office cited only the English member. At trial, the defendants put in the entire Nemelka family, and the courts treated the family (not Lake alone) as "exhibit[ing] completely the state of the art at the time the Gray patents were taken out."

Reference B — U.S. Patent No. 209,829 (Nemelka) — the single most relevant U.S. prior-art patent

• Full citation: US 209,829 A, "Improvement in roller-mills for reducing grain," Lorenz Nemelka, of Simmering, near Vienna, Austria.
• Dates: application filed 11 May 1878; patent dated 12 November 1878. (This is the U.S. member the Supreme Court refers to as "a patent in the United States, November 12, 1878.")
• Brief description: A mill with a fixed middle roller (e) and two adjustable rollers (d, f). The bearings of the adjustable rollers sit in swinging brackets (i) pivoted at (k) on eccentrics (l), the eccentrics being formed in one piece with the axles (k); displacing levers (m) keyed to the axles adjusts the rollers toward/away from roller (e). Set screws (n) prevent the adjustable rollers from jamming against the fixed roller. Pressure is self-acting by the weight of the swinging rollers (the roller axial lines lie slightly above the line o–k), so the rollers "will give way and be lifted if an unusually hard substance should come between them." Levers (m) carry set screws (p) engaging slots in standards (q) to lock the levers. Claims include the combination of roller d or f, brackets i, axle k, eccentrics l, and levers m.
• § 102 mapping — potential anticipation of:
  • Claim 3 (roll on upright swinging arms whose lower ends are mounted on vertically adjustable pivots, the pivots "serving both to sustain and adjust the rolls"): Nemelka's swinging brackets i on eccentrics l plus set screws n read directly on "sustain and adjust."
  • Claim 7 (roller + upright supporting arm D + eccentric sleeve F on which the arm's lower end is mounted + screw E securing the sleeve against rotation): Nemelka's eccentric l + brackets i + locking set screws p/n is the closest prior art; the distinction is eccentric-sleeve-on-the-pivot-bolt (Gray) vs. eccentric-formed-on-the-axle (Nemelka).
  • Claim 1 (roll falls back by gravity; pivots in advance of the roll axis): Nemelka relies on self-acting gravity action of the swinging rollers, but in the opposite direction (gravity closes, hardness opens). This is a partial, not clean, § 102 read on claim 1.
  • Claim 4 arguably: Nemelka is described by the trial court as showing "a spring to hold the movable roll to its grinding position and pressure, with means for regulating the pressure of the spring and the grinding distance by means of cams, screws, and nuts." But note the conflict — the Supreme Court describes the Nemelka buffer as an india-rubber buffer at the lower end of a long descending arm, i.e., not an outside spring on a horizontal rod connecting the two bearings. See § 6 below.
• Source: https://patents.google.com/patent/[US209829A](/patent/US209829A)/en ; https://patentimages.storage.googleapis.com/21/57/46/bdafc225114f06/[US209829](/patent/US209829).pdf

Reference C — Nemelka Austrian patents

• Full citation: Austrian patents to Lorenz Nemelka, dated 15 January 1875 and 22 May 1875.
• Brief description (as found by the courts): vertical adjustment by a screw worked by a worm — held by the trial court to be "the equivalent of Gray's cam or eccentric"; a swinging frame with a cam on the pivot for separating rolls without disturbing parallelism; sliding pivoted attachment for parallelism; spring with cam/screw/nut regulation of pressure and grinding distance.
• § 102 mapping: Claims 2 and 3 (vertically adjustable pivot; horizontal separation without disturbing adjustment). Foreign patents of 1875 are § 102(a)/(b) art (and, ante-dating Gray's application by ~4 years, effectively unassailable on date). The courts did not treat them as fully anticipating the granted claims.

Reference D — Nemelka French patent of 1875

• Full citation: French patent to Lorenz Nemelka, 23 June 1875. A working model of the French patent was in evidence.
• Brief description: As with the Austrian, plus the feature emphasized in Barnard & Leas: rods (Gray's rods G analogue) moved inwardly/outwardly by a cam, the rolls being thrown apart without disturbing the grinding adjustment — there accomplished by cams acting on the pivots of the swinging frame (connected on a shaft for simultaneous operation) rather than on transverse rod ends.
• § 102 mapping: Claim 5 (adjusting rod with a stop at one end and stop/holding devices at the other) is the target here — the French patent is the record's closest showing of a rod-and-cam spreading device, but the structure differs (cam on the pivot vs. stops on the rod). Claim 4 as well.

Reference E — "Die Mühle"

• Full citation: An exhibit/publication referred to in the record as Die Mühle (the German milling journal of that name).
• Date: not established in the sources I retrieved — I cannot give you a date, and I will not guess one.
• Brief description: "shows very plainly a spring arrangement similarly located [at the lower end of a long descending arm of the movable bearing] by which the movable roll is made to yield to a sudden pressure."
• § 102 mapping: § 102(a)/(b) printed publication aimed at claim 4 (spring urging the bearing, with tension adjustment) and the spring aspect of claim 5. This is the only non-patent prior art identified in the record, and its date is the weak link.

3. Other prior-art references named in the trial record (not the Patent Office citation)

These appear in the defendants' answer and the opinions in the companion Gray-patent litigation:

Do not read the following as prior art to 222,895. They are co-asserted later patents from the same 1895 suit, and they post-date Gray's July 1879 filing, so they cannot anticipate it: Gray No. 238,677 (8 March 1881) — held void for want of novelty; Odell reissue No. 10,139 (20 June 1882); Birkholz No. 269,623 (26 December 1882). Likewise the Dowling/Marmon prior-art set — Caine Nos. 78,423 and 137,051; T. J. West No. 100,573; H. E. Keeler No. 254,140; Gray's prior patent No. 228,525 — was cited against other patents in the same suit, not against 222,895. I include them only so the record is complete.

4. Google Patents "Similar Documents" — these are not citations (important)

The Similar Documents list on US222895A is a modern algorithmic similarity list of later patents, not a citation record. All but two post-date the patent, so they cannot be prior art:

• Pre-date 23 Dec 1879 (only two, and both in unrelated arts): US 193,289, 17 July 1877 ("Improvement in sta[v]mp-mills" — OCR garbled); US 211,187, 7 January 1879 (leather-splitting machine).
• Post-date it, therefore incapable of being prior art, though listed: US 274,210 (1883), US 255,715 (1882), US 255,859 (1882), US 280,992 (1883), US 288,886 (1883), US 294,418 (1884), US 367,570 (1887), US 401,243 (1889), US 420,026 (1890), US 421,996 (1890), US 434,028 (1890), US 440,157 (1890), US 444,348 / 444,349 / 444,350 (1891), US 571,266 (1896), US 667,777 (1901), US 881,287 (1908), US 893,713 (1908), US 988,293 (1911), US 1,065,998 (1913), and CN 102596413 B (2014).

Flagging a possible confusion risk: two prior-art patents from other cases share numbers with the patent at issue in this analysis. US 4,222,895 (Hockerson) and US 10,221,780 (Champion Power Equipment) are different patents. Per the operating rule I have not "corrected" any number and have excluded them.

5. Consolidated § 102 mapping — most relevant prior art per claim

6. Bottom line

1. Most relevant prior art: the Nemelka family, and above all U.S. Patent No. 209,829 (filed 11 May 1878; patented 12 November 1878) read together with the English patent to Lake, No. 3,328 of 1877 (dated 28 February 1878). The Lake patent was the only citation of record in prosecution; US 209,829 is the only U.S. prior-art patent in the family and the closest structural match (swinging brackets on eccentrics with locking screws — Gray's claims 2, 3 and 7).
2. Highest-confidence § 102(b) reference on date: US 209,829 and Lake 3,3328/1877 — both more than one year before the July 1879 filing. The Nemelka Austrian (15 Jan and 22 May 1875) and French (23 June 1875) patents are older still.
3. Critical legal caveat: as adjudicated, no reference ever invalidated claims 1–7 of 222,895. In The Roller Mill Patent, 156 U.S. 261 (1895), the Court held the patent not infringed and expressly held it "not a pioneer patent," construing claims 4–6 narrowly on the strength of Gray's prosecution amendments; only the companion patent 238,677 was held void for want of novelty. Judge Blodgett's contrary findings on the Nemelka art (Barnard & Leas, 43 F. 527) were never converted into a judgment of invalidity as to 222,895.
4. The § 102 case against claims 5 and 6 is effectively nonexistent in the retrieved record. The strongest attacks are on claims 2, 3 and 7.

7. Confidence and limitations

• High confidence: US 209,829's number, inventor (Lorenz Nemelka, Simmering), title, filing date (11 May 1878), issue date (12 Nov 1878), and content; the Lake English patent No. 3,328 of 1877 and its 28 Feb 1878 date; the Nemelka Austrian (15 Jan / 22 May 1875) and French (23 June 1875) patents; that Lake was the sole citation in the 14 Aug 1879 rejection; that claims 4–6 were the asserted claims; the outcome of 156 U.S. 261.
• Moderate / flagged uncertainty: the exact July 1879 filing day (OCR "July so, 1879"); the date and nature of the "Die Mühle" exhibit (I could not date it); the "Dingler" and "Schacht" references' dates and whether patent or machine; Nagel & Kaemp and Schacht patent numbers (not stated in the opinions retrieved); whether any of the pre-1879 "Similar Documents" were ever actually cited in prosecution (I found no evidence they were).
• Negative findings (explicit): No "References Cited" or "Cited By" record exists for US222895A on Google Patents, and I could not locate any USPTO-hosted citation record for it; no modern (post-1900) litigation, IPR, reissue or reexamination involving US 222,895 was found.
• Corrections/additions to the earlier sections of this analysis: (i) the previous Litigation Summary listed five proceedings; the record also shows Consolidated Roller-Mill Co. v. Richmond City Mill-Works, 40 F. 474 (E.D. Mich., 9 Nov 1889) (Gresham, J.; provisional injunction denied on an undertaking) — a sixth proceeding. (ii) The Nemelka U.S. patent number (209,829) was not identified earlier and is now supplied. (iii) The earlier summary's inference that the application was filed 30 July 1879 stands as inferred, not confirmed — nothing in the new searches contradicts it.

I'll verify the two pre-filing-date references in the "Similar Documents" table, plus the prior art of record that the litigation identified, before running the § 103 analysis.

Strong results. Let me pin down the two pre-filing-date references in the "Similar Documents" table and confirm the Nemelka US patent's specifics.

Obviousness Analysis — US 222,895 (Gray) under 35 U.S.C. § 103

0. Two threshold corrections before the analysis

(a) § 103 did not exist when this patent issued. US 222,895 issued December 23, 1879 on a July 1879 application. Section 103 was enacted by the Patent Act of 1952. The governing test in 1879 was the judge-made "invention" requirement of Hotchkiss v. Greenwood, 52 U.S. 248 (1851) — whether the improvement required "more ingenuity and skill … than were possessed by an ordinary mechanic acquainted with the business." That test is functionally § 103's ancestor but is not identical: it has no Graham factors, no KSR "predictable variations" doctrine, and no developed secondary-considerations framework. This matters because the 19th-century opinions discussed below sometimes conflate anticipation (§ 102) with want of invention — which is exactly why they read so much like § 103 analysis. Treat everything below as a modern reconstruction, not as what the examiner or the courts actually applied.

(b) What the "Prior Art section of this page" actually contains — read literally. The Google Patents page for US222895 gives three things:

That third point is decisive and is frequently mishandled. Of the 25 "similar documents," only two predate the filing date:

• US 193,289 (1877-07-17) — OCR title on the page: "Improvement in staivip-fvhlls" (garbled; I could not retrieve the document and expressly decline to guess what field it is in or what it discloses).
• US 211,187 (1879-01-07) — "Improvement in leather-splitting machines."

Every other entry postdates the filing date and is therefore not statutory prior art against this patent: US 571,266 (1896); US 667,777 (1901); US 1,945,853 (1934); US 444,349 / 444,350 / 444,348 (1891); US 434,028 (1890); US 294,418 (1884); US 421,996 (1890); US 988,293 (1911); US 274,210 (1883); US 401,243 (1889); US 255,859 (1882); US 420,026 (1890); US 288,886 (1883); US 1,065,998 (1913); US 280,992 (1883); US 255,715 (1882); US 367,570 (1887); US 881,287 (1908); US 440,157 (1890); US 893,713 (1908); and CN 102596413 B (2014). A US document could still qualify as of its filing date (pre-AIA § 102(e)), but a filing date earlier than July 1879 for any of these is not credible and I have not verified any of them.

My searches for US 193,289 and US 211,187 returned no primary documents. I found only unrelated leather-splitting art in the same search neighborhood (US 149,542 Stearns 1874; US 147,172 Reynolds 1874 — see § 3 below). So the page's own "prior art" table, taken literally and date-filtered, supplies essentially no usable § 103 art.

1. The prior art that actually matters (from the record, verified)

The real art is not on the Google page — it is the Nemelka/Lake family, and it is unusually well documented because the Supreme Court described it.

US 209,829's own words (verified from the specification PDF):

• "the bearings of the two rollers d and f are arranged in swinging brackets i, pivoted at k on the eccentrics l, which are formed of one piece with the axles … by displacing the levers m m … the rollers d and f may be adjusted nearer to or farther from the roller e."
• "To prevent the possibility of the adjustable rollers … being jammed to the non-adjustable roller e, their swinging brackets i's rest on the set-screws n n."
• "The central axial lines of the rollers d and f lie a little above the straight lines o k, drawn from the center of the axial line of the rollers to the center of the axles k … this pressure is self-acting, being caused by the weight of the swinging rollers … the said rollers d and f will give way and be lifted if an unusually hard substance should come between them."

Judge Blodgett's findings in Consolidated Roller-Mill Co. v. Barnard & Leas Mfg. Co., 43 F. 527 (N.D. Ill. 1890), which is the closest thing to a § 103 holding in the record:

"Nemelka also showed a swinging frame carrying a movable roller, with a cam working upon the pivot by which the swinging frame was fastened to the fixed frame, by means of which the rolls could be separated without disturbing their parallelism, and a provision for adjusting the rollers to parallelism by sliding the pivoted attachment upon the fixed frame. He also shows a spring to hold the movable roll to its grinding position and pressure, with means for regulating the pressure of the spring and the grinding distance by means of cams, screws, and nuts … I cannot resist the conclusion that all which Gray did … was to secure the same adjustments … by, in many respects, the same instrumentality, but differently located, or well-known equivalents of such instrumentalities."
— https://law.resource.org/pub/us/case/reporter/F/0043/0043.f.0527.1.pdf

Judge Blodgett also expressly held the screw-worked-by-a-worm to be "the equivalent of Gray's cam or eccentric." That is a § 103 conclusion written in 1880s vocabulary.

2. Level of ordinary skill and the four objectives

A PHOSITA in 1879 is a mill engineer/millwright in a mature art actively displacing millstones with rolls. Both Gray's specification and Judge Blodgett identify the same four desiderata:

1. Vertical adjustment (tram) — axis in a common horizontal plane.
2. Horizontal adjustment (parallelism) — surfaces parallel, gap constant.
3. Separation at rest — rolls held apart when not grinding, "without disturbing parallelism or vertical adjustment," to avoid wedging/binding.
4. Spring pressure — hold rolls to grinding pressure, yield to hard foreign matter.

Nemelka supplies all four. That is the whole ballgame for motivation.

3. Claim-by-claim obviousness

4. The named combinations and the motivation to combine

Combination A (primary, claims 1–3, 7): US 209,829 (Nemelka) alone, or in the alternative Lake, English No. 3,328 of 1877 (as representative of the Nemelka family) in view of the ordinary skill in the art.

• Same field of endeavor — both are roller mills for reducing grain.
• Motivation — Nemelka addresses the identical four objectives; where Gray differs, it is by relocating a known instrumentality ("differently located") or by substituting a mechanical equivalent (worm-screw ↔ eccentric). For claim 1, the motivation is the porcelain-roll/wedging problem Gray himself recites.

Combination B (claims 4–6): US 209,829 or Lake (spring-pressed movable roll with cam/screw/nut pressure and gap adjustment, and a separation mechanism that does not disturb parallelism) in view of common mechanical hardware — a threaded tie rod with an adjustable nut, a jam-nut, and a positive shoulder stop.

• Motivation — to make grinding pressure adjustable and repeatable, and to protect rolls from tramp metal; both references already address both needs. Reasonable expectation of success: high — every element is a purely mechanical, predictable load-path component.

Combination C (claims 2, 3, 6 — supplementary, from the same search trail, not from the page's table): the leather-splitting roll art — US 149,542 (Stearns, Apr. 7, 1874) (springs "fitted under the blocks, and thus afford a yielding action" to a roller) and US 147,172 (Reynolds, Feb. 3, 1874) (adjustable collars "to prevent the rings … from wearing unevenly into the rubber roll"). These are analogous if the problem is framed as "maintaining uniform roll pressure and compensating uneven roll wear," and both predate July 1879. But this is a weak combination and I flag it as such: for a 19th-century court these are different fields of endeavor, and the leather-splitting patents do not address tram or parallelism.

5. What a challenger must survive

1. Prosecution history cuts both ways. The Commissioner's August 14, 1879 letter (the Supreme Court opinion also uses "August 19" — a discrepancy in the record itself) rejected Gray's broad original claims over English patent 3,328 of 1877. Gray then inserted "peculiar" and "special" into the statement of invention and withdrew all five original claims, substituting the seven narrower ones. A challenger relying on Nemelka/Lake is on ground the examiner already plowed — good for the challenger. But the granted claims were allowed over that art, which the decider must weigh.
2. Coombs is the principal rebuttal. In Consolidated Roller-Mill Co. v. Coombs, 39 F. 25 (E.D. Mich. 1889), the court upheld the validity of claims 2, 3, 4, and 5 against the Lake/Nemelka anticipation defense and found infringement. That is a contemporaneous judicial finding that the Nemelka family does not dispose of these claims. A § 103 case built only on Nemelka/Lake collides head-on with Coombs.
3. The invalidity record is thinner than the non-infringement record. Note carefully: the Supreme Court's holding was no infringement of claims 4, 5, and 6 of 222,895, and the companion Gray 238,677 was void for want of novelty. The Court did not invalidate 222,895. Judge Blodgett likewise held 222,895 "not infringed" (and void only as to 238,677). The summary section's framing of claims 4–6 as a non-infringement holding is correct and should not be overstated into an invalidity holding.
4. The patent is admittedly not a pioneer. The Supreme Court characterized the invention as a narrow combination, so no liberal-equivalents indulgence — but that cuts for the challenger on obviousness while cutting against the challenger on a broad reading of claims 4–6. The two doctrines pull in opposite directions.
5. Secondary considerations, if pleaded. A sustained multi-forum litigation campaign by Consolidated Roller-Mill Co. (Wisconsin, Michigan, Illinois) over a 17-year term is weak but non-trivial evidence of commercial value; the specification's identification of the porcelain-roll problem is evidence of a recognized need. In 1879 practice these carried far less weight than they would today.

6. Bottom line

• Strongest § 103 attack: a single-reference attack on claims 1 and 2 (and derivatively 3 and 7) over US 209,829 (Nemelka) — same field, same four objectives, same swinging-bracket-on-eccentric architecture, and an express teaching that the position of the roll axis relative to the pivot determines the direction of the gravity-induced force. Coombs stands in the way of claims 2, 3, 4, 5.
• Strongest combination attack: US 209,829 / Lake + ordinary threaded-rod-and-nut hardware, against claims 4, 5, and 6, motivated by adjustable and repeatable grinding pressure plus yield to tramp material. This is the position Judge Blodgett essentially took in 43 F. 527 ("well-known equivalents of such instrumentalities").
• Weakest claim to attack: claim 1's precise direction of gravity bias — Gray inverted Nemelka's self-acting closing pressure into a self-acting opening bias. That inversion is a small but genuinely purposive step tied to a stated problem (porcelain rolls, grain wedging at rest), and it is the one feature no reference in the record literally shows.

Confidence: high on the Nemelka/Lake content, US 209,829's disclosure and dates, and Blodgett's findings (all verified this session from primary/reporter sources). Low/unknown: the disclosures of US 193,289 and US 211,187 — the only two date-qualifying entries in the page's "Similar Documents" table — for which my searches returned no primary document; I make no representation about what either discloses. Also flagged: the August 14 vs. August 19 letter-date discrepancy inside the Supreme Court's own opinion, and the 1877/1878 dating of the English Lake patent.