[857]*857JOHN R. BROWN, Circuit Judge.
On this appeal by the Lessors, we are confronted with the correctness of the Judge’s charge as given to the jury, his failure to give certain issues and instructions requested by the Lessors, and whether the Judge erred in granting the Lessees’ motion for partial summary judgment. These questions primarily involve a construction of an oil and gas lease to determine if the well involved was “producing gas only” within the meaning of the lease, and if so, whether shut-in gas royalty payments were timely made in the manner authorized by the lease. Although as to much of the case we find no error and to that extent affirm the judgment, we conclude that the case must be reversed and remanded for a partial new trial on the limited issues later discussed.
This is a suit brought by the Lessors 1 to declare an oil and gas lease terminated, to remove it as a cloud on their title, and for damages. Originally brought in the State Courts of Texas in the form of a statutory trespass to try title action, the Lessees2 removed it to the United States District Court. 28 U.S.C.A. § 1441. The case was tried and submitted to a jury under special interrogatories. F.R.Civ. P. 49. The jury having answered the interrogatories favorably to the Lessees, judgment was entered in their favor, and the Lessors have appealed.
The lease involved in this case was executed on March 26, 1947, by predecessors of the Lessors. The lease was an ordinary oil and gas lease for a primary term of 10 years, which would expire March 26, 1957. It contained the following habendum clause:
“2. Subject to the other provisions herein contained, this lease shall remain in force for a term of ten years from this date, called primary term, and as long thereafter as oil, gas or other mineral is produced from said land, or as long thereafter as Lessee shall conduct drilling or reworking operations thereon with no cessation of more than sixty consecutive days until production results, and if production results, so long as any such mineral is produced.”
No question is here presented of the payment of delay rentals during the primary term, it being undisputed that the lease was in full force and effect on December 20, 1956, when Lessees began drilling operations on the land. These operations continued until the well began to flow gas on April 26, 1957. The Lessees permitted the flowing to continue until 2:00 p. m. on April 30. During this period, the Lessees tested the well, obtained complete records as to tubing and casing pressures and rate of flow, and measured the amount of gas, oil, drilling mud, kerosene and water produced from the well. At 2:00 p. m., the well was shut in so that a 48-hour shut-in pressure test could be run. This test was taken at midnight on May 2, 1957. On May 6, the Lessees displaced the kerosene in the well with water and began removing the tubing and packer which had been inserted previously. Gas was vented for a short time on May 7. The “Christmas tree” and other surface equipment was installed on May 8. The well was swabbed on May 9 and much of the actual drilling equipment was removed. From May 9 at 4:00 p. m. until midnight on May 13, the well was allowed to flow. An employee of Lessees read the gauges, inserted the “Y” tube and made notations on his records every eight hours. These activities continued until midnight on May 13 at which time the well was capped and shut-in because of a lack of an accessible market for gas. The Lessees stipulated that as of that time the well was then capable of producing gas in paying quantities. It remained capped until August 16, 1960, at which time it was connected to a pipeline and gas was produced and marketed. To keep the lease alive, the Lessees paid [858]*858annual shut-in gas royalties.3 The outcome of the case turns on whether this was permissible and timely done.
At the outset it is good to re-emphasize that under Erie, we are sitting as a Texas court applying Texas law as best we can divine it, conscious all the while that our decision loses authority when the first Texas writing court declares to the contrary. Smoot v. State Farm Mutual Ins. Co., 5 Cir., 1962, 299 F.2d 525. It is thus incumbent upon us to apply the principles, fashioned by the Texas courts, of oil and gas law to the facts of this case.
In Texas, the ordinary oil and gas lease is a transfer and conveyance of real property so that title to the oil and gas in place is vested in the lessee thereunder. Waggoner Estate v. Sigler Oil Co., 1929, 118 Tex. 509, 19 S.W.2d 27; Brown, Oil & Gas Leases § 3.02 (1958). However, unlike the mineral deed, it is not an absolute fee simple. Nor is it an estate on condition subsequent requiring the re-entry of the grant- or. It is a determinable fee leaving the lessor a possibility of reverter or of reacquiring the absolute fee simple title, less whatever minerals in the meantime are produced and marketed. Texas Co. v. Davis, 1923, 113 Tex. 321, 254 S.W. 304, 255 S.W. 601; Sullivan, Oil & Gas Law 95 (1959). It is a fee simple title because it may continue forever. And yet it is determinable because it may come to an end upon the happening of certain contingencies. Upon the happening of these contingencies or limitations, there is no forfeiture, since the estate which was granted has merely run its course. It has expired automatically under its own terms. Vernon v. Union Oil Co., 5 Cir., 1959, 270 F.2d 441; Haby v. Stanolind Oil & Gas Co., 5 Cir., 1955, 228 F.2d 298; Stephens County v. Mid-Kansas Oil & Gas Co., 1923, 113 Tex. 160, 254 S.W. 290, 29 A.L.R. 566; Caruthers v. Leonard, Tex.Com.App., 1923, 254 S.W. 779. The terms are ordinarily spelled out in the lease’s habendum clause, a clause of special limitation which marks the duration of the estate granted.4 The effect of this clause has been summed up in the classic statement of Professor A. W. Walker, Jr.:
“Neither unavoidable delays or accidents, acts of God, unfavorable economic conditions, nor financial difficulties of the lessee will afford an excuse for the failure to comply literally with the provisions of this clause in the absence of an express stipulation otherwise contained in the lease.”5
In this approach, the presence and the content of, and compliance with, provisions extending the primary term are of critical significance. Thus, if a lessee completes a gas well during the primary term of the lease but the lease form does not contain a shut-in gas royalty clause, the lease will automatically expire and the lessee’s interest will terminate and revert to the lessors upon the expiration of the primary term unless oil, gas or other minerals are then actually being produced in paying quantities. Stanolind Oil & Gas Co. v. Barnhill, Tex.Civ.App., 1937, 107 S.W.2d 746, error ref’d. Likewise, under a lease containing a shut-in gas royalty clause, if a lessee were to complete a commercially productive gas well during the primary term but the well is (and remains) shut-in for lack of market, the lease will automati[859]*859cally terminate upon the expiration of the primary term if the lessee fails to tender or pay the shut-in gas royalty prescribed by the lease prior to the expiration of the primary term. Freeman v. Magnolia Petroleum Co., 1943, 141 Tex. 274, 171 S.W.2d 339. However, in the event the lessee completed a gas well after the expiration of the primary term and the well is capped or shut-in immediately for lack of a market, the question of whether the lessee has any additional time within which to tender shut-in gas royalty depends on the provisions of the lease as construed by the Texas courts. Gulf Oil Corp. v. Reid, 1960, 161 Tex. 51, 337 S.W.2d 267; Skelly Oil Co. v. Harris, 1962, Tex., 352 S.W.2d 950.
Since it is undisputed that on neither May 2, nor May 13, 1957, was there actual production of gas in paying quantities, it is clear that the determinable fee in the present lease automatically expired of its own accord at the end of the ten-year primary term (March 26, 1957) unless some provision such as the continuous drilling clause operated to extend the terms of the grant.
The Lessors contend that the trial Judge erred in sustaining Lessees’ motion to limit the issues for trial and alternatively for partial summary judgment. In effect the Judge’s order did two things. First, it determined as a matter of law that if this well were one “producing gas only,” then the manner and time of payment of shut-in gas royalty beginning May 13, 1957, and thereafter in 1958, 1959 and 1960 fully complied with the lease. Second, it specified that . the only fact issue for trial was whether the well was one “producing gas only.” As to the first we conclude that the District Court was correct and affirm such holding. The second element is tied into the question of whether productivity of gas only relates to the well (a) as actually completed, or (b) as it should have been completed.
It simplifies matters to dispose of the first element even though that assumes the well was a gas producer only. For if the payments were not timely, there is no need for further inquiry or retrial. The Lessors strongly rely on the case of Gulf Oil Corp. v. Reid, 1960, 161 Tex. 51, 337 S.W.2d 267, 39 Texas L.Rev. 519, as foreclosing the issue of whether shut-in royalties were timely paid. In that case the primary term of the lease expired on December 9, 1948. The lessee began the drilling of a well a few days before the expiration of the primary term and continued drilling operations up to and including January 18, 1949. The well was capable of producing gas in paying quantities on the date of its completion, but it was immediately capped and shut in due to the lack of a market. The lessee did not tender the shut-in gas royalty payment to the lessor until a month later on February 19, 1949. The well was not connected and gas was not produced and sold until November 22,1949. The Texas Supreme Court held that the 60-day clause 6 in the lease did not give the lessee 60 days from the capping of the well in which to tender shut-in royalties.
Although our case involves a Reid-type well,7 the lease involved here is not a Reid-type lease. Mindful that in an area of vast economic importance to Texas in which, unlike the Texas courts, Erie, Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188, does not afford us the ready means of correcting our mistakes, we think it best that— without more — we trim our sails as close[860]*860ly as possible to the rationale developed in authoritative Texas decisions. We may, of course, point out as one obvious distinction between this case and Reid that the lease in Reid gave the lessee 60 days between the “abandonment of one well and the commencement of drilling or reworking operations on another until production is obtained.” Since there was no abandonment of one well and the commencement of another in Reid, the Court could conclude that the 60-day clause was inapplicable. But whatever might be the elucidation of distinction, we think that our case is directly controlled by the very recent Texas Supreme Court case of Skelly Oil Co. v. Harris, 1962, Tex., 352 S.W.2d 950. The facts in that case were almost identical with those involved here and the lease contained substantially the same provisions.8 Drilling started prior to the end of the primary term. The well was completed as a gas well after the term expired. But gas was not then produced. The well was temporarily capped until the lessee could get connections with a pipe line. This connection was made and actual production commenced some 41 days following completion of the well. In an opinion written by Justice Walker with his characteristic skill, expressly rejecting an argument made by the lessor 9 which is substantially the same argument made by the Lessors here, the Court held that production within the 60-day period, following the completion of the well kept the lease alive. In other words, during this 60 days neither actual nor constructive “production”' was required.
The Lessors here insist that Harris is distinguishable because there actual production was obtained. We do not think the cases are distinguishable on these grounds. The habendum clause of the present lease opened with this phrase, “Subject to the other provisions herein contained * * *.” It was thus required to yield to any and all other provisions which would affect the duration of the lease. In this connection, Paragraph 310 of the lease provided in part that upon the completion of a well capable-of producing gas only, the lessee had an option. He could either begin actual production or he could pay shut-in gas royalties. Upon the payment of the shut-in royalty “it will be considered that gas is-being produced within the meaning of [the habendum clause].” Under its plain terms, the lease could be extended by either actual or constructive production. Of course constructive production through shut-in royalty payments is. available only where the well is capable of actual production.11 See Kidd v. Hog-gett, Tex.Civ.App., 1959, 331 S.W.2d 515, error ref’d, n. r. e. But there is no problem as to this since this physical capability as of May 13, 1957, was stipulated.
[861]*861With Harris as our Erie point of departure in the construction of this typically Texas contract, we think it had this meaning. At the expiration of primary term, the lease shall not terminate if the lessee is then engaged in drilling or reworking operations. However, if lessee ceases such drilling or reworking operations, the lease shall terminate unless within 60 days (1) he commences drilling, reworking operations, or (2) production results. This interpretation is consistent with the primary purpose of the lease, i. e., to achieve production. This clause allows the lessee to maintain the lease (and thus possibly achieve production) by the diligent prosecution of operations then being conducted at the expiration of the primary term. Rogers v. Osborn, 1953, 152 Tex. 540, 261 S.W.2d 311; Stanolind Oil & Gas Co. v. Newman Bros. Drilling Co., 1957, 157 Tex. 489, 305 S.W.2d 169.
Thus, from the time the well was capped on May 13, the Lessees had 60 days (to July 12,1957) in which to begin “production,” actual or constructive. The question therefore remains: was there a timely payment within that period in 1957, and in succeeding years on or before the vicarious anniversary date ?
The shut-in royalty clause does not set out in detail the manner and method of payment of shut-in gas royalty. The delay rental clause of the lease (Paragraph 4) expressly provides that the Lessee may pay or tender delay rentals by Lessee’s cheek or draft mailed to the Lessor at the specified address or by deposit to the Lessor’s credit in the named depository bank. Thus, the delay rental clause expressly authorized (1) the payment by Lessee’s check or draft and (2) tender of rentals by mail. However, the shut-in royalty clause does not so authorize payments for shut-in royalty to the Lessor and the lease here does not contain a provision, which apparently is common in numerous lease forms, to the effect that the shut-in royalty may be paid “in like manner and amount” as delay rentals. Prom this, the Lessors would have us hold that all shut-in gas royalty under the lease should have been paid directly to Lessors in United States currency.
We may accept Lessors’ contention that in the absence of express agreement, the delivery of a check to the payee is merely a conditional payment which does not become absolute until the check is actually paid by the drawee bank.12 A check is simply a written order on a bank, executory in nature, instructing the bank to make certain payment, and so far as the payee’s rights against the drawee are concerned, this order may be revoked by the drawer at any time before the check is paid or in some manner accepted by the bank on which it is drawn.13 We are further in agreement with the Lessors’ argument that the act of making the check on May 13 and mailing it to Lessors on that date did not at that time have the same legal effect that would result from the act of tendering $25 in currency to each of the two Lessors on May 13.
However, we do not regard this as being decisive under Texas law. The Lessors do not go far enough. Conceding that a check is merely conditional payment, once the check has been deposited or cashed, and makes the complete commercial cycle back to the drawee bank where it is finally accepted and paid, such payment relates back to the time the check was delivered to the payee. Muldrow v. Texas Frozen Foods, 1957, 157 Tex. 39, 299 S.W.2d 275; Texas Mutual Life Ins. Ass’n v. Tolbert, 1940, 134 Tex. 419, 136 S.W.2d 584. This is a sensible-result, especially considering that histori[862]*862cally the Iáw of negotiable instruments is an outgrowth of the Law Merchant in which law, far more than accommodating itself to business necessities, was rather the product of those business practices. No harm lurks in this blending of law and the needs of modern day commerce. If the check is dishonored on presentment to the drawee, no timely “payment” has been made. If, as was true here, the checks are paid, they had at the end of the cycle the same marketability for full cash value that the payee-Lessors considered they had.
Under our holding in this case, it is immaterial whether the 1957 shut-in royalty payment was delivered to Lessors prior to or after May 13. For the check was made, delivered, deposited and paid by the drawee bank well within the additional 60-day period following May 2 or May 13 as allowed by the lease. Actually it was deposited on May 16, 1957. However, an anniversary date must be assigned to the well in order to determine the timeliness of the 1958, 1959 and 1960 tenders. Assuming that the anniversary date for these subsequent payments of shut-in royalty is determined from the date on which the well was actually completed, no genuine fact issues were raised which required this to be tried. The operations conducted at the well site between May 2 and May 13 related to the testing and completion of the well. Until it was capped on that date, it was not completed for purposes of fixing an anniversary date. If, on the other hand, the date for subsequent years is the anniversary of the time of the receipt of the 1957 shut-in payment, it was not earlier than May 14. On either assumption the subsequent payments were timely made. Each of the payments for 1958, 1959 and 1960 was made by check dated April 22 of the respective year, and each was transmitted by mail to Lessors shortly thereafter. The 1958 check was deposited by Lessors on April 24. The 1959 check was deposited on May 11, and the 1960 check was deposited on May 13. It is also undisputed that each of the checks was honored and paid in the regular course of business by the drawee bank. Thus, it is clear that each of these three checks was received and deposited on or before May 13 of the respective year and was timely.
Neither do we think there is merit in the contention that the payments were ineffective because $25 was not tendered separately to each of the two Lessors. The checks, though payable jointly to the two Lessors bore endorsements of both, and the proceeds were deposited one-half ($25) to the account of each man. The Lessors, father and son, were the joint owners of the land. They were paid royalties thereon jointly. It was uncontradicted that the receipt, endorsement, and deposit of each check by one was authorized by the other. We find no error in the District Court’s holding that the time and manner of payment of shut-in royalty complied with the lease as a matter of law. Limited summary judgment was therefore proper.
But as this assumed the well was one “producing gas only,” we must now turn to the question of whether this assumption was well founded. Otherwise constructive production by payment of shut-in royalty would not keep the lease alive. We must here distinguish between the well as (a) actually completed and (b) as it should have been completed. As to (a), we have no doubts that the well as acUially completed was a well “producing gas only.” The jury so found in answer to the special interrogatories, and there is ample evidence to support their finding.
In this connection, until Texas declares to the contrary we are bound by Vernon v. Union Oil Co., 5 Cir., 1959, 270 F.2d 441, 38 Texas L.Rev. 807. In that case, this Court, sitting as a Texas court, held that a well capable of producing liquid condensate in paying quantities was not a “gas only” well. We emphasized, however, that this was to be judged, not as a mere theoretical capability of the well producing some liquid hydrocarbons such as condensate. Rather, it was to be determined in the light of [863]*863reasonable practical standards, 1. e., whether “ * * * it would be reasonable to operate the well to produce liquid condensate alone * * *,’’ 270 F.2d 441, 446. Whether a well is capable of producing condensate (or some other liquid “non-gas” products) in paying quantities may well involve inquiry into such things as the expenses necessarily and reasonably incurred by virtue of obligatory conservation statutes and Texas Railroad Commission Orders.14 As flaring is prohibited,15 only two choices are open: (1) recycle and extract condensates, or perhaps (2) store the gas underground. Choice No. (2) merely substitutes one natural storage for another. And, so long as the “well” is confined as the Court’s instructions did to the well as actually completed, the evidence amply justifies the jury verdict that it was not economically feasible to attempt to recover liquid hydrocarbons from this gas through choice No. (1) by recycling. Indeed, as we understand Lessors’ brief, this is virtually conceded.
In doing so, however, Lessors by no means abandon this hard fought litigation as a juridical dry hole. Rather, they say, this demonstrates the seriousness and the error of the trial Court’s purposeful limiting of the jury to the well as (a) actually completed, rather than (b) as it should have been completed.
The question is sharply posed by adequate objections preserved below and pressed vigorously here as to the following instruction given to the jury:
“You are instructed in connection with each and all of the foregoing Special Issues that in deciding said issues you will deal with the well in question as it was actually completed and not as it could have been or might have been completed.”16
The determination of the correctness of this instruction encompasses all of Lessors’ additionally assigned errors concerning the propriety of the Judge’s refusal to give certain requested instructions and issues dealing with the well as it might or could have been completed.17
The well was drilled to a total depth of 9,965 feet at an approximate cost of $225,000. The Tonkawa formation, from which the well is presently producing, is located at a depth between 6,595 feet and 6,658 feet. In the completion of the well, the lower portion was pinched off and cemented. The well as completed as a gas well was perforated at 6,611-15 feet. It is Lessors’ contention that the well should have been completed with perforations somewhere between 6,615-30 feet. It is their theory that had this been done, the well would have been completed as a Vernon gas well — one also producing liquids in paying quantities— in which event constructive production through payment of shut-in royalty does not keep the lease alive.
[864]*864Contrary to the contention made in Lessees’ supplemental post-argument briefs that this is a wholly new theory hatched in the exigencies of oral argument, the record is otherwise. Lessors’ formal pleadings assert three distinctive theories.18 The first of these is that the shut-in royalties were not timely and properly paid. The second is that the well as completed is capable of producing something more than “gas only” and this negatives the use of the constructive production provision. The third theory, advanced alternatively, is that the well reasonably could have been completed as one capable of producing liquids in paying quantities, and that Lessees knew of this when the well was perforated to produce gas only. Even more important, the trial was a hard-no-holds-barred-battle of competing fact, quasi-fact and expert witnesses on whether it was, or was not, prudent to try to bring this in as an “oil” well.
In support of the action of the trial Court in confining jury consideration to the well as actually completed, either one •or both of two bases might be urged. First would be the theory that even though the known facts demonstrate that an “oil” well could be brought in, a lessee Las the right by any appropriate engineering methods to deliberately complete it so it will produce gas only thereby setting in motion a lease of unlimited duration so long as the nominal shut-in gas royalty payments are made. The second, and most likely the approach followed by the trial Judge, is that the evidence was not sufficient to show a reasonable likelihood that a prudent operator •could or should have brought it in as an “oil” well.
While it is evident as to this that the •case from the Lessors’ point of view is not a strong one and, on the contrary, the technical evidence from the Lessees’ emphasizes many reasons why such a course was not practicable, we conclude that there was sufficient evidence to require jury submission and determination.
No purpose would be served in discussing this evidence in detail. It suffices to summarize it briefly. On two drill stem tests oil was discovered at 6,610-30 and 6,620-50. On the basis of this, Schlum-berger electric logs, and other recited facts which the jury could credit, a qualified expert expressed the firm opinion that had perforations been made at “ * * * the 6620 to 6630 interval * * * ” in all “reasonable probability” the recovery “would have been oil.” Equally important, at least two inter office telegrams reflected a precise, strong recommendation by Lessees’ field superintendent having immediate responsibility for the well that it be perforated at 6614-30 feet. Though sent and received during the crucial days preceding completion operations, these recommendations were rejected by remote, but higher, headquarters. This may have taken on the color of a purposeful plan to avoid making an “oil” well of it, as there was also testimony of a statement made by an agent of Lessees that an oil well was about to be brought in, and this would shake up the home office which wanted a gas well. As bearing on what ought reasonably to have been ascertainable in 1957, there was also evidence as to a well some two miles away completed several years later as an oil producer in sands correlated to 6621-31 feet in the gas well in controversy.
It is true, of course, that there is no word testimony, as such, which speaks in terms of commercially paying quantities. But with such pieces as the amount of oil recovery on the drill stem tests, the strong recommendation of the field superintendent and other evidence, there was enough to permit an inference of commercial feasibility.
[865]*865The evidence was by no means one-sided. Indeed, Lessees made a formidable showing. Prior to- the drilling of the well in suit, several wells had been drilled to the Tonkawa formation in surrounding areas. Of these, only two were producers and they were producing gas. The remainder had been abandoned as salt water wells or “dry holes.” As stated above, by a well completed in the same horizon some four years later and nearly two miles away, the Lessors undertook to show that the instant well should have been completed as an oil well. However, at the time of trial, this other well was producing salt water at a rate of three times as much as oil produced. Experts predicted that the well could not long continue as an oil producer. There was also evidence that the shaly portions of the Tonkawa formation under the well in suit would not create an effective barrier against salt water encroachment. In fact, at the time of trial', the very well in suit was producing salt water in an amount equal to the condensate recovered —and this was true even though it was completed as a gas, not an oil, well.
Although the Judge ostensibly overruled Lessees’ motion for directed verdict, the effect of giving the qualifying instruction and refusing to give Lessors’ requested issues and instructions (or suitable modifications of them) was a partial granting of such motion. The net result was that the jury was not allowed to pass on one of Lessors’ principal theories. Having exercised his sound discretion in submitting the case to the jury on special interrogatories, F.R.Civ. P. 49(a), the Judge was bound to submit the controlling issues raised by the pleadings and the evidence. Travelers Ins. Co. v. Truitt, 5 Cir., 1960, 280 F.2d 784; Clegg v. Hardware Mutual Casualty Co., 5 Cir., 1959, 264 F.2d 152; Maryland Casualty Co. v. Broadway, 5 Cir., 1940, 110 F.2d 357. We think that on the record before us, the theory reflected by the issues and instructions requested by Lessors was adequately raised by the pleadings and the evidence, and the trial Court erred in refusing to submit this phase of the case to the jury under appropriate instructions. As a necessary corollary, this means also that the Judge erred in giving the instruction which limited the jury’s consideration to the well as actually completed. This necessitates a new trial.
The new trial, however, should be a partial one. F.R.Civ.P. 59 (a), 28 U.S.C.A. § 2106. The question for retrial, broadly stated, is whether under all the pertinent circumstances disclosed by the evidence on the retrial, the well in all reasonable probability would have been completed as an “oil” or at least a Vernon-gas-oil well either in the range of 6615 to 6632, or perhaps 6611 to 6632 feet and a diligent, prudent operator acting in good faith would have undertaken to complete it as such, rather than as a gas-only well at 6611-6615 feet. If found favorably to Lessors, judgment would be entered cancelling the lease.19 Vernon v. Union Oil Co., supra. If, on the other hand, the jury finds either that (a) “oil” in paying quantities would not have been recovered or (b) a prudent [866]*866operator would not have completed it as an “oil” well in the elevation of 6611 to 6632 feet, judgment would be entered for Lessees. It bears repeating that if the effect of the jury verdict is that a prudent operator would have completed the well as it was actually completed at 6611-6615 feet as a gas well, the Lessees are entitled to judgment. For reasons discussed at length, the findings on the first trial that the well as actually completed produced gas only is affirmed and is not to be retried. Neither is the time nor manner of payment of shut-in royalty.
We think it appropriate to sound a caveat. We do not intend to predict now what the outcome of the retrial should, or may, be. All we hold is that this evidence raises this issue. It is futile to anticipate what the evidence may be on the retrial. It is almost certain to be different, and its sufficiency inevitably is a matter for initial determination by the trial Judge applying the principles here laid down but without any artificial effort to match that evidence, bit by bit, against that contained in the present record. See Smoot v. State Farm Mut. Auto. Ins. Co., 5 Cir., 1962, 299 F.2d 525, 534.
We would also emphasize that we are not dealing with the right of the lessee ordinarily to choose between several possible strata in which to complete a well.20 Nor would we in the least imply that there is a duty to complete in an horizon thought most advantageous to a lessor even though, by reason of salt water or otherwise, it is doubtful that oil in paying quantities would be recovered. We are, rather, dealing with completion in the same immediate formation. If the facts warrant a finding that there is a substantial probability that oil can be produced in paying quantities,21 the lessee will not be allowed to maintain the lease through constructive production by pinching off commercially productive sands to bring in a gas well only.
Affirmed in part, reversed and remanded in part.