Claiming for Division: Why Patent Prosecutors Should Sometimes Draft Patent Claims to Invite Restriction

Executive Summary: This article argues that patent prosecutors should sometimes treat a restriction requirement as a commercial tool rather than a prosecution inconvenience. For platform technologies—such as a life-sciences program with claims to a compound, formulation, manufacturing method, dosing regimen, biomarker, drug-device combination, and second indication—the business value of the patent estate may lie in preserving separately protectable assets that can later support financing, licensing, enforcement, divestiture, or product-line expansion. Where the specification genuinely supports multiple independent or distinct inventions, drafting claims to make those divisions visible can invite a proper USPTO restriction requirement, creating a record that may support true divisionals and the § 121 safe harbor against obviousness-type double patenting. The strategy is valuable because the alternative may be a continuation-heavy family dependent on terminal disclaimers, which can shorten term, limit alienability, and condition enforceability on continuing common ownership. The article cautions that this approach works only if counsel preserves the statutory and prosecution predicates: the applications must be true divisionals filed as a result of the restriction, consonance must be maintained, divisional timing must be managed before issuance, and the restriction response must be drafted as both an election paper and a future diligence or litigation exhibit.

I.             Introduction

Patent strategy often fails not because the first patent is weak, but because the patent family is too indivisible. A platform technology rarely has one commercial center of gravity: value may reside in a compound, formulation, manufacturing method, dosing regimen, biomarker, delivery device, software layer, or later-discovered use. A claim strategy that forces those assets into one prosecution path may protect the first product while leaving the business with fewer options for licensing, financing, divestiture, enforcement, or future development.

That is the commercial case for inviting restriction. A restriction requirement is usually treated as an inconvenience because it forces election, withdraws non-elected claims from immediate examination, and often requires later divisional filings. In the right case, however, those effects can be useful: a properly supported restriction requirement can help convert a broad disclosure into a divisional architecture that maps to commercial control points.

This article argues for a deliberately commercial approach to restriction practice. Where the specification genuinely discloses multiple commercially meaningful, supportable, independent or distinct inventions, prosecutors should sometimes draft claims so that the separateness of those inventions is visible enough for the Patent Office to require restriction. The strategy is not to manufacture artificial claim groups or procedural burden. It is to align prosecution architecture with enterprise value.

A Working Hypothetical

Consider a platform life-sciences company developing a small-molecule therapy. Its original disclosure supports claims to a lead compound or genus, a bioavailability-enhancing formulation, a manufacturing route, a dosing regimen, a biomarker-selected patient population, a drug-device combination, and a second therapeutic indication. At filing, the compound may appear to be the commercial center of gravity. Five years later, clinical data may make the dosing regimen central, regulatory feedback may elevate the formulation, a partner may want the drug-device combination, a manufacturer may care most about the process claims, and a second indication may become the growth asset.

A one-track claim strategy can obscure those separate value drivers. A restriction-aware strategy instead presents them as distinct, supported claim families from the beginning. If the examiner requires restriction, the resulting groups can become more than procedural categories. They can become the working map for later divisionals, diligence, licensing, and enforcement. The hypothetical is used below to make the doctrine concrete.

II.          Restriction Requirements Are Not Merely Administrative Events

The starting point is statutory. Section 121 provides that, “[i]f two or more independent and distinct inventions are claimed in one application,” the Director may require the application to be restricted to one of the inventions. 35 U.S.C. § 121. The implementing rule similarly provides that, when two or more independent and distinct inventions are claimed in a single application, the examiner will require the applicant to elect one invention for examination; claims to non-elected inventions that are not canceled are withdrawn, subject to reinstatement if the restriction is later withdrawn or overruled. 37 C.F.R. § 1.142(a)–(b).

Proper restriction, however, requires more than an examiner’s conclusory statement that the claims are different. The MPEP instructs that restriction is proper only when the inventions are independent or distinct and there would be a serious search or examination burden if they were examined together. MPEP § 803. MPEP § 808 likewise requires the examiner to identify reasons why the inventions are independent or distinct and why examining them together would impose a serious burden. MPEP § 808.01.

Those requirements are often treated as obstacles to be overcome. They should also be treated as drafting opportunities. In the life-sciences hypothetical, a claim set that separately presents the compound, formulation, manufacturing route, dosing regimen, biomarker-defined treatment method, drug-device combination, and second indication may give the examiner a principled basis to require restriction. The different groups may implicate different search fields, different prior-art universes, different statutory issues, and different technical problems.

The prosecutor’s strategic insight should be this: the restriction requirement is the Office’s own recognition that the application contains more than one independent or distinct invention. That recognition may later matter far more than the temporary inconvenience of election.

III.      The § 121 Safe Harbor: Doctrinal Conditions

The argument for drafting claims to invite restriction depends on a paradox: the § 121 safe harbor is valuable precisely because it is narrow. It is not a general immunity from obviousness-type double patenting. It is not a family-wide blessing of every claim later pursued in a continuation chain. And it is not an equitable clean-up doctrine available whenever a patent owner can tell a sympathetic restriction story after the fact. It is a statutory prohibition with defined predicates.

Section 121 provides that, when two or more independent and distinct inventions are claimed in one application, the Director may require restriction to one invention. 35 U.S.C. § 121. It then provides that a patent issuing on an application in which a restriction requirement has been made, or on an application filed as a result of that requirement, “shall not be used as a reference” against a divisional application, the original application, or a patent issuing on either, if the divisional application is filed before issuance of the patent on the other application. Id. The safe harbor therefore operates as a prohibition on using one restricted-family patent as an obviousness-type double-patenting reference against another qualifying application or patent. It does not itself declare that the claims are patentably distinct. It prevents the patentee from being penalized for complying with an Office-imposed division. See Applied Materials, Inc. v. Advanced Semiconductor Materials Am., Inc., 98 F.3d 1563, 1568 (Fed. Cir. 1996); Studiengesellschaft Kohle mbH v. N. Petrochemical Co., 784 F.2d 351, 358–59 (Fed. Cir. 1986) (Newman, J., concurring).

That statutory structure explains both the attraction and the danger of the strategy. A restriction requirement can create an unusually strong prosecution record: the Office itself has required the applicant to divide claims because the claims are directed to independent or distinct inventions. But because § 121 can prevent an otherwise available double-patenting challenge, the Federal Circuit applies it strictly. See Geneva Pharms., Inc. v. GlaxoSmithKline PLC, 349 F.3d 1373, 1382 (Fed. Cir. 2003); G.D. Searle LLC v. Lupin Pharms., Inc., 790 F.3d 1349, 1354 (Fed. Cir. 2015). The doctrinal conditions are formal: Office-compelled division, divisional status, filing as a result of the restriction requirement, consonance, and statutory timing.

A. The safe harbor requires Office-compelled division, not voluntary portfolio slicing

The first requirement is a real restriction requirement. Section 121 is triggered by the Office’s determination that the application claims independent and distinct inventions. 35 U.S.C. § 121. A voluntary decision by the applicant to split claims among multiple applications is not the same thing. The safe harbor exists because the applicant has been required by the Office to divide claims; it is not a reward for unilateral portfolio management.

That distinction is central to the hypothetical. The life-sciences company could file continuations directed to the formulation, biomarker, or second indication for ordinary portfolio reasons. But if the original application presents those inventions distinctly enough for the Office to require restriction, later divisionals may have a statutory safe-harbor story that voluntary continuations lack. The applicant cannot command the Office to restrict, but it can draft claims that make a legitimate restriction visible. MPEP § 804.01; see also Bristol-Myers Squibb Co. v. Pharmachemie B.V., 361 F.3d 1343, 1347–48 (Fed. Cir. 2004).

That is why the phrase “provoke a restriction requirement” should be understood carefully. The applicant cannot command the Office to restrict. Nor should the applicant manufacture false separateness. But where the disclosure genuinely supports multiple independent or distinct inventions, the claims can be drafted so that the Office has a clear and principled basis to require restriction. The safe harbor begins with that Office action.

B. The later application must be a true divisional

The second requirement is equally unforgiving: the application invoking the safe harbor must be a divisional, or at least must descend from a proper divisional in a way that preserves the statutory requirements. Section 121 uses the term “divisional application.” 35 U.S.C. § 121. The Federal Circuit has taken that language seriously.

Pfizer is the leading warning. There, the original application contained compound, composition, and method-of-use claims. The examiner imposed a restriction requirement. Pfizer filed a divisional that matured into one patent and a continuation-in-part that matured into another. The district court applied § 121, but the Federal Circuit reversed, holding that the safe harbor is limited to divisional applications and patents issuing from such applications. Pfizer Inc. v. Teva Pharms. USA, Inc., 518 F.3d 1353, 1362 (Fed. Cir. 2008). A continuation-in-part was not good enough. Id. at 1362–63.

The reason is not formalism for its own sake. A divisional is carved out of an earlier application and does not add new matter. A continuation-in-part, by definition, adds new matter. If CIPs could claim § 121 protection, the safe harbor could shield subject matter that was never part of the restricted claims in the original application. That would convert a protection against Office-compelled division into a broader immunity for later-added inventions. Pfizer refused to take that step. Id. at 1361–63.

Amgen added a further discipline. The Federal Circuit recognized that a patent need not always issue directly from the first divisional application to receive § 121 protection; intervening continuations may not destroy safe-harbor protection if they descend from a proper divisional filed as a result of a restriction requirement. But the court refused to treat applications filed and designated as continuations as if they were divisionals where they did not descend from a divisional. Amgen Inc. v. F. Hoffmann-La Roche Ltd., 580 F.3d 1340, 1353–54 (Fed. Cir. 2009). In other words, a continuation may sometimes live within a protected divisional chain, but an ordinary continuation is not converted into a divisional merely because the applicant later wishes it had been one.

For prosecution purposes, this is a drafting and docketing point as much as a legal point. If the claims are directed to non-elected subject matter and § 121 protection may matter, the application should be filed and identified as a divisional. In the hypothetical, if the biomarker group is restricted out and later becomes commercially important, the follow-on application should be a true divisional, not a generic continuation tree to be recharacterized later. The specification should not add new matter, and the priority claim, application data sheet, preliminary amendment, and remarks should be consistent with divisional status.

C. The divisional must be filed “as a result of” the restriction requirement

The third requirement is causation. It is not enough that a restriction requirement exists somewhere in the family. The application invoking § 121 must have been filed “as a result of” that requirement. 35 U.S.C. § 121; Bristol-Myers, 361 F.3d at 1347–48.

That requirement has real bite. In Bristol-Myers, the Federal Circuit held that the patentee could invoke § 121 only if the divisional application was filed as a result of the restriction requirement and was consonant with that requirement. 361 F.3d at 1348. The court rejected a loose relationship between the restriction history and the later patent. Id. The restriction requirement must be part of the causal and legal lineage of the later divisional.

Boehringer refined the point. The Federal Circuit held that § 121 can apply to a divisional of a divisional, assuming the other requirements are satisfied. Boehringer Ingelheim Int’l GmbH v. Barr Labs., Inc., 592 F.3d 1340, 1352 (Fed. Cir. 2010). That holding is important for sophisticated portfolio planning. The safe harbor is not necessarily limited to the first divisional filed after the restriction requirement. A properly maintained divisional chain can preserve the benefit. But Boehringer also emphasized that the “as a result of” requirement must be satisfied by both the challenged patent and the reference patent. Id. at 1351–52. It is therefore not enough to show that one family member traces to the restriction. The relevant patents must share the required relationship to the same restriction architecture.

This point should influence how prosecutors draft and maintain family charts. A useful chart should not merely list priority claims and expiration dates. It should identify the application in which the restriction requirement issued, the group covering the compound, the group covering the formulation, the group covering the biomarker method, the group covering the drug-device combination, and each divisional that pursued those groups. The question is not simply “Are these patents related?” The question is whether both the challenged patent and the asserted reference patent can be traced, without crossing the restriction lines, to the same Office-imposed division.

D. Consonance is the price of the safe harbor

The fourth requirement is consonance. Consonance requires that the line of demarcation between the independent and distinct inventions identified by the examiner be maintained in the later applications and patents. Gerber Garment Tech., Inc. v. Lectra Sys., Inc., 916 F.2d 683, 688 (Fed. Cir. 1990); Symbol Techs., Inc. v. Opticon, Inc., 935 F.2d 1569, 1579 (Fed. Cir. 1991). If that line is crossed, the safe harbor does not apply. Symbol, 935 F.2d at 1579.

This is the most important litigation-facing consequence of restriction-aware drafting. If the examiner divides the hypothetical claims into Group I for the compound, Group II for the formulation, Group III for manufacturing, Group IV for biomarker-guided treatment, and Group V for a drug-device combination, later applications should not casually mix those groups. A claim that is commercially attractive but straddles the compound and biomarker groups may create more long-term risk than value.

St. Jude is the cautionary centerpiece. The Federal Circuit explained that the safe harbor arose because restriction requirements and double-patenting challenges can otherwise collide unfairly. St. Jude Med., Inc. v. Access Closure, Inc., 729 F.3d 1369, 1376–77 (Fed. Cir. 2013). But the court refused to apply the safe harbor because consonance had not been maintained. Id. at 1381–82. The problem was not merely the challenged patent in isolation; the court examined the challenged patent, the reference patent, and the restricted patent to determine whether the family had honored the restriction lines. Id. at 1378–82. Because the sibling patent overlapped the restricted subject matter, the safe harbor failed. Id. at 1381–82.

That point is critical for prosecutors. Consonance is not a one-application inquiry. It is a family-architecture inquiry. A later divisional may look clean when viewed alone, but safe-harbor protection can still be lost if another patent in the family crosses the examiner’s demarcation line and is then used as a double-patenting reference. The prosecutor must therefore police not only the claims in the divisional being filed, but also the claims in sibling applications, continuation applications, and patents that may later become references.

For this reason, restriction-aware drafting should be accompanied by restriction-aware claim amendment. During prosecution, amendments often migrate toward commercially important embodiments. That is natural. But when § 121 protection is a strategic objective, every amendment should be checked against the original restriction group. A claim that is commercially attractive but straddles two restricted inventions may create more long-term risk than value.

E. Section 121 protects against a reference; it does not cure indistinct claims outside the statute

A further subtlety is that § 121 does not make indistinct claims distinct. It prevents a qualifying patent from being used as a reference in an obviousness-type double-patenting rejection or challenge. 35 U.S.C. § 121. If the statutory requirements are not met, the ordinary double-patenting analysis returns.

This matters because prosecutors sometimes speak of the safe harbor as if it “solves” double patenting across a family. That is too broad. The safe harbor may prevent a particular patent from serving as a double-patenting reference against a qualifying divisional or original application. But if the later claims fall outside the restriction architecture, if the application is not a true divisional, if the relevant application was not filed as a result of the restriction requirement, or if consonance has been lost, § 121 will not rescue claims that are otherwise not patentably distinct. See Pfizer, 518 F.3d at 1362–63; St. Jude, 729 F.3d at 1381–82; G.D. Searle, 790 F.3d at 1354–56.

IV.       Preserving the Safe Harbor in Prosecution

The doctrine has an immediate prosecution consequence: the safe harbor must be built while the applications are pending. The following points are practice disciplines rather than additional black-letter predicates, but they determine whether the statutory conditions can later be proved.

A. Build the safe harbor before issuance, not in litigation

The Federal Circuit has been particularly skeptical of efforts to reconstruct § 121 protection after issuance. G.D. Searle and Janssen make this point sharply.

In G.D. Searle, the patentee attempted to use reissue to correct a patent-family problem. The original patent had issued from an application that was prosecuted as a continuation-in-part rather than as a divisional. In reissue, the patentee deleted new matter and redesignated the application as a divisional. The Federal Circuit held that the safe harbor did not apply. G.D. Searle, 790 F.3d at 1354–56. Deleting new matter in reissue did not retroactively alter the nature of the application from which the patent had issued. Id. at 1355. The court also found that the relevant patents did not share the required common lineage in the divisional chain. Id. at 1356.

Janssen reached the same practical conclusion in the reexamination context. There, the patent owner attempted to bring a challenged patent within § 121 by amending the patent during reexamination to remove new matter and redesignate the application as a divisional. The Federal Circuit rejected the attempt, holding that a patent owner cannot retroactively bring a challenged patent within the safe harbor by amendment in reexamination. In re Janssen Biotech, Inc., 880 F.3d 1315, 1321–23 (Fed. Cir. 2018). The court stated that, at the latest, the application must be properly designated as a divisional by the time the challenged patent issues. Id. at 1323.

Safe-harbor protection is not something to be fixed in district court, reissue, or reexamination. It must be engineered while the applications are pending. The practitioner who wants the benefit of § 121 should make the record look like a § 121 record from the beginning: original claims to multiple inventions; an Office-imposed restriction; a clean election response; a divisional application directed to non-elected subject matter; no new matter; claim amendments that maintain consonance; and a family chart documenting the lineage.

B. Let the filing-before-issuance condition drive divisional timing

Section 121 also contains a timing condition: the divisional must be filed before issuance of the patent on the other application. 35 U.S.C. § 121. In practice, this reinforces a simple prosecution discipline: do not wait until after the parent or another relevant family member issues to decide whether the non-elected invention matters.

The safest practice is to make divisional decisions before allowance and certainly before issuance of the relevant related patent. In the hypothetical, counsel should use notice of allowance in the elected case as a checkpoint: identify the non-elected groups, evaluate clinical data, partner discussions, manufacturing scale-up, and second-indication plans, and decide which divisionals to file while the statutory path remains available. A restriction requirement that lapses into issuance without timely divisional action may become a missed opportunity rather than a portfolio asset.

C. Favor clean architecture over after-the-fact characterization

The Federal Circuit cases reward clean architecture and punish clever reconstruction. Applied Materials and Symbol show that § 121 can protect patents where the applicant has respected the Office-imposed division. Applied Materials, 98 F.3d at 1568–69; Symbol, 935 F.2d at 1579–80. Boehringer shows that the protection can extend beyond the first divisional where the chain remains tied to the restriction requirement and the other statutory requirements are met. 592 F.3d at 1352. But Pfizer, Amgen, St. Jude, G.D. Searle, and Janssen show that the court will not rescue a family that is structured as something other than a divisional chain and later relabeled when double patenting becomes a litigation problem. Pfizer, 518 F.3d at 1362–63; Amgen, 580 F.3d at 1353–54; St. Jude, 729 F.3d at 1381–82; G.D. Searle, 790 F.3d at 1354–56; Janssen, 880 F.3d at 1321–23.

That case law should be framed as a practical drafting mandate. If the prosecutor wants the benefit of restriction, the application must be drafted and prosecuted as if the restriction record will later be scrutinized by a district court and the Federal Circuit. It is not enough to say that the applicant could have filed a divisional. It is not enough to say that the later claims correspond roughly to subject matter that was restricted out. It is not enough to remove new matter later. The family must be structured so that the statutory story is true.

V.          Terminal Disclaimers Are Not a Cost-Free Substitute

One common response to obviousness-type double patenting is that the applicant can simply file a terminal disclaimer. That response is too casual. Terminal disclaimers may solve a prosecution problem, but they can create a long-term asset problem.

A terminal disclaimer filed to overcome nonstatutory double patenting does more than disclaim patent term. Under 37 C.F.R. § 1.321(c)(3), it must include a provision that the patent will be enforceable only for and during the period in which it is commonly owned with the application or patent that formed the basis for the double-patenting rejection. 37 C.F.R. § 1.321(c)(3). The regulation also provides that terminal disclaimers are binding on successors and assigns. 37 C.F.R. § 1.321(b).

That is a critical transactional point. The requirement is best described as common ownership, not “joint ownership” in the colloquial sense. For the enforceable life of the terminally disclaimed patent, enforceability depends on common ownership with the reference patent or application. In the hypothetical, that may matter if the company later wants to assign the manufacturing process to one partner, license the biomarker claims by field, or spin out the drug-device combination.

The Federal Circuit’s decision in SimpleAir, Inc. v. Google LLC illustrates the point. The patents there had terminal disclaimers filed during prosecution to overcome obviousness-type double-patenting rejections, and those disclaimers required expiration with, and common ownership alongside, the relevant parent patents or applications. 884 F.3d 1160, 1166–67 (Fed. Cir. 2018). The court also observed that terminal disclaimers can affect both patent term and alienability. Id. at 1167–68.

The ownership problem is even more acute in collaborative research, university licensing, joint-development, and divestiture settings. In In re Hubbell, the Federal Circuit held that common ownership is not a prerequisite for an obviousness-type double-patenting rejection. 709 F.3d 1140, 1147–49 (Fed. Cir. 2013). That matters because a terminal disclaimer may not be practically or legally available when ownership has already diverged. See id.; see also In re Fallaux, 564 F.3d 1313, 1318–19 (Fed. Cir. 2009) (discussing double-patenting concerns in the context of divided ownership).

Restriction-aware drafting responds to that problem at the front end. A restriction-supported divisional path may reduce the need to rely on terminal disclaimers later. That does not mean § 121 eliminates all double-patenting risk. It means that a properly preserved safe-harbor position may be preferable to a prosecution strategy that solves every double-patenting rejection by tying the family together through terminal disclaimers.

VI.       The ODP Landscape Makes Family Architecture More Important, Not Less

Obviousness-type double patenting remains a live and sometimes counterintuitive risk in modern patent families. The Federal Circuit’s post-URAA cases show that expiration dates, patent term adjustment, filing order, issuance order, and claim overlap can interact in unexpected ways.

In Gilead Sciences, Inc. v. Natco Pharma Ltd., the court held that, under the circumstances presented there, a later-issued but earlier-expiring patent could serve as an obviousness-type double-patenting reference against an earlier-issued but later-expiring patent. 753 F.3d 1208, 1215–17 (Fed. Cir. 2014). In AbbVie Inc. v. Mathilda & Terence Kennedy Institute of Rheumatology Trust, the court likewise emphasized that obviousness-type double patenting remains concerned with improper timewise extension of exclusivity. 764 F.3d 1366, 1373–74 (Fed. Cir. 2014).

More recently, In re Cellect, LLC held that obviousness-type double patenting for a patent that received patent term adjustment is assessed using the expiration date after PTA has been added. 81 F.4th 1216, 1228–31 (Fed. Cir. 2023). The court also rejected a late, after-expiration terminal-disclaimer theory as a cure. Id. at 1231–32.

At the same time, Cellect should not be overstated. In Allergan USA, Inc. v. MSN Laboratories Private Ltd., the Federal Circuit clarified that a first-filed, first-issued, later-expiring claim was not invalid for obviousness-type double patenting over a later-filed, later-issued, earlier-expiring claim sharing a common priority date. 111 F.4th 1358, 1368–70 (Fed. Cir. 2024).

Taken together, these cases do not create a single mechanical rule for every patent family. They create a planning imperative. For the hypothetical company, a family covering a compound, formulation, dosing regimen, manufacturing route, biomarker method, and second indication may have different prosecution histories and different expiration profiles. Prosecutors should not assume that common priority, continuation practice, common ownership, or terminal disclaimers will make ODP disappear. Where the disclosure supports it, restriction-supported divisionals may provide a cleaner architecture.

VII.   The Commercial Case for Inviting Restriction

The commercial value of restriction-aware drafting is clearest when the invention is not a single point improvement, but a platform. The life-sciences hypothetical illustrates the point. A patent family that treats the compound, formulation, manufacturing route, dosing regimen, biomarker method, drug-device combination, and second indication as separate claim families may be more useful than a family that treats them as mere fallback embodiments of one invention.

A. Optionality and staged investment

Platform companies rarely know at filing which asset will become commercially dominant. The elected claims may be pursued first because they are closest to the launch product or least exposed to prior art. The non-elected claims can remain candidates for divisionals as clinical data, regulatory feedback, financing, partner interest, and competitor behavior clarify value.

In the hypothetical, the compound may justify the first election. But if a dosing regimen later becomes essential to the label, if the formulation becomes the differentiator, or if the biomarker-selected population becomes the reimbursement story, a preserved divisional path can be more valuable than the marginal savings from a narrower original claim set. The point is not to prosecute everything aggressively at once. It is to preserve the right to make later decisions with better information.

That staging should be tied to milestones. At allowance of the elected case, counsel should revisit each non-elected group and ask whether it maps to a current product, a likely design-around, a partner opportunity, a regulatory development, or a future financing story. Patent budgets should be measured against the value of the options preserved, not simply against the number of applications filed.

B. Diligence and transaction readiness

A useful restriction requirement often does what a good patent-portfolio manager would do anyway: it separates assets into commercially intelligible categories. In the hypothetical, one group may correspond to the core compound, another to manufacturing, another to treatment methods, another to the biomarker population, and another to the drug-device combination. If later divisionals track those groups, the family is easier to understand in diligence.

Investors and acquirers do not merely ask whether a company has patents. They ask whether the patents cover the product, likely design-arounds, future indications, manufacturing control points, and partner-facing assets. They also ask whether divisional rights remain pending, whether ODP risk is controlled, and whether terminal disclaimers impair separability. A family organized around a clean restriction requirement can answer those questions more persuasively than a family built through ad hoc continuations.

This is also where record discipline becomes commercial. Buyers and licensees will ask whether the restriction groups were clear, whether later applications tracked them, and whether the family was built as a true divisional architecture rather than retrofitted after the fact. See G.D. Searle LLC v. Lupin Pharms., Inc., 790 F.3d 1349, 1354–56 (Fed. Cir. 2015); In re Janssen Biotech, Inc., 880 F.3d 1315, 1321–23 (Fed. Cir. 2018). The file history should make clear what was restricted, what was elected, what remained non-elected, whether the applicant traversed, and how later divisionals map to the examiner’s groups.

C. Licensing and enforcement leverage

Restriction-aware drafting can produce a more usable licensing package. A licensor with one broad patent often has a blunt instrument. A licensor with a structured family can license more precisely: manufacturing claims to a process partner, biomarker claims to a diagnostic partner, drug-device claims to a device collaborator, and second-indication claims to a therapeutic licensee. That granularity can support field-of-use licenses, supply-chain licenses, defensive cross-licenses, and settlement structures.

The enforcement benefits are similar. A single patent may be vulnerable to a narrow noninfringement position or a single invalidity challenge. A structured family can present different claim types against different participants in the commercial chain: product claims against manufacturers, method claims against users or service providers, manufacturing claims against suppliers, and device-combination claims against integrated platform operators.

This does not mean more patents are always better. Poorly differentiated patents can create ODP problems, terminal-disclaimer dependencies, and claim-preclusion complications. The point is more precise: a family built around genuinely different restricted inventions can provide different claims directed to different commercial chokepoints, while preserving a cleaner explanation for why those patents exist separately.

D. Transactional flexibility

The restriction strategy is especially attractive where future asset separation is plausible. Universities may license different fields of use to different companies. Platform start-ups may spin out non-core applications. Pharmaceutical companies may sell manufacturing technology while retaining therapeutic claims. A patent family encumbered by terminal disclaimers can make those transactions harder because enforceability may depend on continuing common ownership.

A restriction-supported divisional family may offer a cleaner alternative. If the original application was drafted to present multiple real inventions, if the Office required restriction, if true divisionals pursued non-elected groups, and if consonance was preserved, the resulting assets may be easier to explain and potentially easier to transact. The family is not free from all ODP risk, but it may have a statutory safe-harbor narrative that does not depend on keeping every asset bundled forever.

The commercial thesis is therefore narrow but important. Inviting restriction is not always desirable. It increases cost, forces election, and may complicate prosecution. But when the business model depends on multiple protectable nodes, the prosecutor should ask whether the claims should be drafted not merely to survive examination, but to create a patent family the business can use.

VIII.                  The Restriction Response Is a Portfolio Document

A restriction requirement should not be answered mechanically. The response may become one of the most important documents in the family history.

The first question is whether to traverse. If the examiner’s grouping is strategically useful, legally sound, and consistent with the applicant’s future divisional plan, non-traversal may make sense. If the examiner’s grouping is overbroad, inaccurate, or likely to create harmful claim-scope implications, traversal may be essential. The decision should not be driven solely by short-term prosecution convenience.

The second question is what to elect. The elected group is often the invention most likely to generate near-term allowance or immediate commercial protection. But that is not always the right answer. The prosecutor should consider product launch timing, competitor design-around risk, expected continuation filings, foreign prosecution, prior-art strength, PTA consequences, and the client’s budget for divisionals.

The third question is what record to create. If the applicant elects without traverse, the response should avoid gratuitous admissions that the non-elected claims are patentably distinct in a way that may later narrow claim scope. If the applicant traverses, the response should preserve the applicant’s disagreement without undermining the ability to file divisionals if the restriction is maintained. If the examiner’s statement of restriction mischaracterizes the invention, the response should correct the record.

That final point has become more important after Focus Products Group International, LLC v. Kartri Sales Co., 156 F.4th 1259 (Fed. Cir. 2025) (see Restriction Requirements as Substantive Limiters: Lessons from Focus Products v. Kartri ). The case should not be read to mean that every election of species narrows claim scope. Its force lies in the later prosecution record: the examiner repeatedly treated certain claims as outside the elected species, the patentee accepted that boundary, and the court held that the patentee could not later reclaim the disclaimed scope. Id. The lesson for restriction-aware prosecution is straightforward: the same record that supports divisional architecture can also narrow claim scope if mishandled.

Restriction-aware drafting therefore requires restriction-aware prosecution. The prosecutor should treat the restriction response as both an election paper and a future litigation exhibit.

IX.       Conclusion

The best patent prosecutors do not draft only for first allowance. They draft for the life of the asset.

A restriction requirement is often inconvenient, but inconvenience is not the same as harm. In the right case, a restriction requirement can be a valuable event: it confirms that the application presents multiple independent or distinct inventions, creates a roadmap for divisionals, supports § 121 safe-harbor arguments, and may reduce the need for terminal disclaimers that sacrifice term and condition enforceability on common ownership.

The strategy requires care. Section 121 is formal. Consonance matters. Divisional status matters. Prosecution responses matter. Terminal disclaimers may still be necessary. And a restriction record can narrow claim scope if mishandled.

But those cautions support, rather than undermine, the central point. If a client’s disclosure is a platform, the claim strategy should not pretend it is a single narrow invention. Where multiple real inventions are disclosed, a sophisticated prosecutor should sometimes draft the claims so that the Patent Office has a principled reason to restrict them. In those cases, provoking a restriction requirement is not gamesmanship. It is disciplined portfolio architecture.