concurring.
I withdraw my opinion of February 26, 1998, and substitute the following in its place.
I joined the Court’s original opinion in this case. On rehearing, it became evident to me that our original opinion, although scholarly,1 went wide of the mark. The rehearing has produced a more accurately focused argument by the parties and consequently a better opinion from this Court. But the plurality insists on writing beyond the problem on points that seriously undermine its own position. Consequently, I can only concur in the Court’s judgment, although I agree with the plurality on rehearing that we cannot reasonably interpret the deed in this case to convey two estates.
The original opinion from this Court was seriously flawed in two respects. First, the Court presupposed that the typical grantor does not intend to make two grants in one deed. Concord Oil Co. v. Pennzoil Exploration, 40 Tex. Sup.Ct. J. 33, 35, 1996 WL 596657 (Tex. 1996). Second, the Court anchored the opinion on a conclusion that the Crosby deed’s “subject to” clause included future leases. 40 Tex. Sup.Ct. J. 33, 38. Pennzoil, on rehearing, could not have been clearer in its attack on these flaws. Surprisingly, Concord, in response, made no effort at all to support the underpinnings of the Court’s opinion. We granted the rehearing and further argument ensued.
Pennzoil rightly contends that whether “most grantors” intend to make two grants in one deed is an irrelevant inquiry unless the Court determines that the document, standing alone, contains internal inconsistencies. See, e.g., Luckel v. White, 819 S.W.2d 459, 462 (Tex.1991). We undermined our opinion by applying our presupposition that a typical grantor intends to convey only one estate before we concluded that the deed was internally inconsistent. Further, we were wrong to conclude that the “subject to” clause of the Crosby deed includes future leases. It does not — at least not clearly. And that is Pennzoil’s second point. Reading the “subject to” clause to include future leases is not implausible, but it is unreasonable. The obligation of the Court is to construe the deed to avoid the disharmony, not create it. Cf National Union Fire Ins. Co. v. CBI Indus., Inc., 907 S.W.2d 517, 521 (Tex.1995). In short, we violated our own rules for construing writings. We read the “subject to” clause in a way that created a conflict within the document, rather than in a way that avoided the conflict.
Today, the plurality undermines its reasoning by continuing to rely on both the improper presupposition about a typical grantor’s intent and the improper reading of the Crosby deed’s “subject to” clause. Concord, by not arguing the point, implicitly concedes that both are flaws in the plurality’s position. Were these the only bases supporting the Court’s judgment, then the dissent’s view should prevail and a different judgment should be rendered. But, there is one other, dispositive reason that dictates the actual judgment in this case. A point that destroys the reading the dissent attempts to give the *464Crosby deed. And a point the plurality merely adds to the Court’s original opinion.
Were the Crosby deed to contain two conveyances rather than one, there would be an unavoidable conflict — a conflict that was, until now, overlooked by us. Even the parties failed to focus on it until oral argument on rehearing. The conflict would arise because, were the granting clause and the “subject to” clause conveying separate estates, they would convey more than Crosby owned. We need only focus on this fact. It is this fact alone that keeps me from joining the dissent, which is otherwise correct. There is nothing inherently wrong with a deed expressing two grants, and where expressed, such grants should be honored. However, the potential for over-grant in the Crosby deed prevents me from concluding that this deed contains two grants.
A proper review of the Crosby deed begins with the four-comers rule, under which we attempt to ascertain the intent of the parties from the language of the deed. See Luckel, 819 S.W.2d at 461. Our job is to harmonize all parts of the deed, if reasonably possible, even though the separate parts of the deed may appear to be contradictory or inconsistent. Cf. National Union, 907 S.W.2d at 520. Importantly, we are not to give the deed a reading that is unreasonable. Id.
We cannot give the Crosby deed the reading that the dissent believes is correct — that the deed makes two conveyances — because that reading is unreasonable. Assuming that the deed makes two conveyances, we would have the granting clause conveying a 1/96 mineral interest. But we would also have the “subject to” clause simultaneously conveying an additional 1/12 (or 8/96) interest in rentals and royalties under the then-current lease. This reading produces an over-grant.
At the time of Crosby’s deed to his grantee, Southland, a mineral lease covered the property. The lessee held title to the mineral estate subject to the possibility that title would revert to Crosby and the other lessors in the future. Crosby, therefore, owned the possibility of a 1/12 mineral interest. Crosby’s reverter interest included the right to royalties under the then-current lease, as do all reverter interests in the absence of language to the contrary. See Luckel, 819 S.W.2d at 464. Therefore, the Crosby deed’s granting clause transferred to Southland 1/96 of Crosby’s reverter interest, carrying with it a corresponding 1/96 share of the royalties due under the then-current lease. See id. If the “subject to” clause were a separate conveyance, it would transfer to Southland an additional 8/96 interest in the royalties due under the then-current lease. Under the dissent’s construction, the granting clause and the “subject to” clause would convey 1/96 plus 8/96 for a total of 9/96 interest in the royalties, a larger interest than Crosby owned. This construction is not reasonable.
As Professor David Pierce has noted in looking at the court of appeals’ opinion in this case, the dissent’s two-grant conclusion makes sense only if one assumes that only one grant operates at a time. See David E. Pierce, Developments in Nonregulatory Oil and Gas Law: The Continuing Search for Analytical Foundations, in 47 Oil & Gas Law & Taxation § 1.05, at 1-24 (1996). That assumption, however, is contrary to the clear language of both the granting and “subject to” clauses.
Because reading the Crosby deed as making two conveyances creates an over-conveyance problem, we may not use that reading as a way of resolving the conflicting fractions in the Crosby deed. Therefore, we must see if there exists another reading that will resolve the conflict between the fractions.
As the plurality notes, the Crosby deed was written in the same year that this Court adopted the Texarkana court of appeals’ opinion in Tipps v. Bodine, 101 S.W.2d 1076 (Tex.Civ.App.—Texarkana 1936, writ ref'd). Tipps blessed the use of differing fractions in granting and “subject to” clauses as the proper way to make a single conveyance when the conveyed interest is subject to an existing lease. See id. at 1079. Construing a deed with differing fractions used similarly to those in the Crosby deed, the Tipps court held that “we know of [no language] that would more clearly and accurately express the intention of the parties” to convey the single larger interest. Id. Simply, Tipps *465supports construction of the Crosby deed to convey all of Crosby’s 8/96 mineral interest.
I do not disagree with the content of the plurality’s dissertation on policy and historical justifications, but the dissertation requires the practitioner to dig to find the nugget that resolves this case. Giving the Crosby deed a reasonable construction, I must conclude that the deed conveyed all of Crosby’s 1/12 mineral interest. Therefore, I concur.
. Patrick H. Martin, Recent Developments in No-nregulatory Oil & Gas Law, in Oil & Gas Law & Taxation § 1.02(2)(e), at 1-15 (1997).