I respectfully dissent. I conclude that the leases in question violate the Crow Allotment Act, as amended, and should be declared void.

The lessees, Scott Land & Livestock Co., et al, urge that the 1948 amendment to the Crow Allotment Act, which deleted an earlier provision prohibiting the renewal of leases of grazing lands prior to one year before their termination, 1 opened the way for the practice now before us. The 1948 amendment, it is said, must be construed consistently with congressional intent gradually to emancipate the Crow Indians from government interference in managing their allotted property.

I am of the opinion that the Act must also be read in the light of the special relationship of the United States to the Indians. The nature of this relationship was aptly first described by Chief Justice Marshall in Cherokee Nation v. Georgia, 30 U.S. (5 Pet.) 1 , 8 L.Ed.2d 25 (1831), as “perhaps, unlike that of any other two people in existence.” 1'd. at 16. He observed that Indian tribes, rather than being foreign states, “may, more correctly, perhaps, be denominated domestic dependent nations” “in a state of pupilage” and concluded that “[tjheir relation to the United States resembles that of a ward to his guardian.” Id. at 17.

This guardian relationship, and the concomitant duty of protection it implies, has been repeatedly acknowledged by Congress and recognized by the courts. As stated by the Supreme Court:

United States v. Kagama, 118 U.S. 375, 384 , 6 S.Ct. 1109 , 30 L.Ed. 228 (1886). Accord, Sunderland v. United States, 266 U.S. 226, 233-34 , 45 S.Ct. 64 , 69 L.Ed. 259 (1924).

The plan of Congress gradually to emancipate the Indians so that they gain greater responsibility over their own affairs is not in conflict with the government’s guardian *1213 ship role. It is, on the contrary, an extension of it. The ultimate aim of guardianship as the term is used here, should be that the wards become, in time, self-sufficient. Legislation meant to further Congress’ plan to emancipate the Indians must therefore be construed in light of the guardian relationship and, necessarily, be limited by it.

I agree that the overall purpose of the 1948 amendment to the Crow Allotment Act was to increase the responsibility of the Crow people over their lands. Legislative history is instructive on the primary means chosen to implement this purpose. The Crow Allotment Act of 1920 2 established a commission to divide the Crow people into two classes, competent and incompetent, with minors belonging to the latter class. No provision was made to reclassify minors when they attained their majority. A later amendment 3 provided that competent Crows could lease their allotted grazing lands for five-year terms without federal supervision. 4 As time passed, however, and minors reached their majority, a smaller and smaller percentage of adult Crows was classified competent. 5 *Congress reversed this trend through the Act of March 3, 1931, 6 authorizing the Secretary to change the classification of any Crow Indian under the Act of June 4, 1920, from incompetent to competent on the recommendation of a committee appointed for that purpose. However, the 1931 legislation failed to define properly the Secretary’s administrative responsibility for proper land use and range management in connection with the leasing of Crow lands and it was not until passage of clarifying legislation in 1948, the amendment now under scrutiny, that the classification of Indians began pursuant to the 1931 Act. It was upon this impending reclassification that congressional attention was primarily focused during enactment of the 1948 amendment. See H.R.Rep.'No.940, 80th Cong., 1st Sess. (1947); S.Rep.No.386, 80th Cong., 1st Sess. (1947).

The 1948 amendment added a provision that any Crow Indian classified as competent should have the full responsibility of obtaining compliance with the terms of any lease made and deleted a provision in an earlier amendment 7 prohibiting the renewal of leases of grazing lands prior to one year before their termination and the renewal of leases of farming lands prior to eighteen months before termination. The practice now in use is that each lease is made for five years but is annually can-celled and renewed for five years. It is argued that, since each five-year lease is within the statutory maximum term and since no prohibition against cancellation of an existing lease and execution of a new lease now appears in the Act as amended, the practice is valid.

It is clear, however, that Congress did not intend, in passing the 1948 amendment, to remove all limitations imposed upon the Crows’ ability to lease their lands. The provision restricting the grazing land leases to five-year terms remained untouched. There is nothing in the legislative history of the Act to show congressional intent to emasculate this provision or to reverse case law construing such restrictions. Indeed, the legislative history is barren as to why the Congress deleted the prohibition forbidding re-leasing until one year before the end of a five-year term. We are left to guess whether the Congress meant that releasing was to be valid during the first four years or forbidden even in the last year. That there was a general attempt to increase the responsibility of Crows over their *1214 land is not a convincing argument for validating all re-leasing practices when considered in light of the guardianship role of the government. Thus, I find that the focal point for deciding the validity of the leasing practice before us is not the absence of the earlier provision in the Act as amended, but the presence of the five-year lease limitation.

It has been frequently recognized that the intent of Congress in placing limitations on the length of leases of trust lands was to shield the Indians from the sharp practices of non-Indians who would despoil them of their property and to protect the Indians from their own lack of business experience. Smith v. McCullough, 270 U.S. 456, 464-65 , 46 S.Ct. 338 , 70 L.Ed. 682 (1926); Sunderland v. United States, supra, 266 U.S. at 233-34 , 45 S.Ct. 64 ; United States v. Haddock, 21 F.2d 165, 167-68 (8th Cir. 1927). Commenting on the reason for lease-length restrictions, the Supreme Court in Smith v. McCullough said:

270 U.S. at 464-65 , 46 S.Ct. at 341 .

The restriction imposed by the Crow Allotment Act is that the term of leases of grazing lands not exceed five years. Under the leasing practice in question, the parties enter into an arrangement whereby the Indian lessors lease their lands for the maximum term, five years. The lessees prepay the rent for the full five years. Thereafter, every year, the Indian lessors and lessees perform a ritual at the office of the lessees whereby the five-year leases are cancelled and new five-year leases, prepared by the non-Indian lessees, are signed. Another year’s rental is paid the Indians.

It is argued that the practice is valid because each lease is only five years long, thus within the letter of the Act’s restriction, that the Indian is in legal possession of his land when he makes each new lease and that he is always free simply to let the lease run to its natural termination without renewing it. I think this argument ignores the economic realities of the situation.

The Indians rely upon United States v. Noble, 237 U.S. 74 , 35 S.Ct. 522 , 59 L.Ed. 844 (1915), where the Supreme Court considered the validity of a leasing practice under a statute, similar to the Crow Allotment Act, which governed leasing of allotted lands to the Quapaw Indians. That statute provided that allottees were authorized to lease their restricted lands for a term not exceeding ten years for mining or business purposes. The lessors would lease the land for ten years but one to four years later would sign a new instrument leasing the land for an additional ten years. The practice created a series of overlapping leases. The Court held that the statute permitted only leases made in possession and that the instruments attempting to lease what was only a reversionary interest were void.

The lessees rely upon United States v. Abrams, 194 F. 82 (8th Cir. 1912), where the court held valid, under the same statute construed in Noble, an outright cancellation of an existing mining lease and the execution of a new mining lease for ten years.

Neither case is controlling on the facts presented here. These leases are not a reversionary interest as in Noble. Each new lease is made while the Indian is in legal possession of his land as was the case in Abrams . But absent in Abrams , and present here, is the practice of prepayment of the original five-year lease and a provision in the cancellation document that nothing therein shall be deemed a release of any claim against the lessor for advances made or rentals prepaid.

Though the courts in Noble and Abrams reached opposite conclusions, the common thread that runs through these cases and *1215 those that follow 8 is the view, implicit or explicit, that Congress intended the Indians to have a free and unencumbered choice, at least once at the end of each statutory-period, to lease their lands to whomever they choose and upon the best terms they can find.

I think the leasing practice here, whereby rentals are prepaid for five years and the obligations of the lessors arising from such prepayments survive annual cancellation of the leases, interferes with the Indians’ right to lease their lands free and unencumbered at least once very five years.

Were the parties businessmen dealing at arm’s length, the practice of prepaying rentals might be thought to benefit the lessors who thereby gain the use of the funds in advance. Here, however, the lessors are Indians whom Congress sought specifically to protect from their own lack of experience in business matters. The lessees are wealthy ranchers. It is not difficult to see how the prepayment practice would work to lock the Indians into a continual cancellation and re-lease procedure. Once the prepaid funds are received and spent, the Indians are faced annually with the alternative of accepting one more year’s rental in exchange for another five-year renewal or going four years without additional income from the land in order to let the lease run out. It has not been shown that the Indians are in an economic position to choose the latter or to cancel the lease and repay the rental sums plus damages to which the lessees may be entitled.

It may be argued that the Indians should be more circumspect in their use of the prepaid funds, retaining them in the event that they should wish to let the leases run out. But this argument misses the mark. It was precisely the Indians’ improvidence in money matters, their inexperience and naivete that spurred Congress to adopt the lease-length restrictions. 9

The five-year lease-length restriction was designed as a fail-safe mechanism to insure that mistakes made by the Indians in leasing their lands would be limited both in consequence and in duration. I find that the leasing practice here, whereby the lessees pay in advance the entire five-year rental, preys on the inexperience and poverty of the Indians and thwarts congressional intent embodied in the Act that the Indians make an unencumbered choice in leasing their property at least once every five years.

It may well be that in a given fact situation, the intent of the statute is being carried out by a particular lease cancellation re-lease practice. It cannot be said that all Indians are forced economically into the lease straitjacket. In fact, it is argued that the Indians have not demonstrated any economic coercion. I feel no initial showing on their part is required. In view of the fact that the purpose of the five-year restriction is to protect the Indians from their own improvidence and from the avarice of non-Indians, I feel the burden of showing that there is no economic coercion is more properly placed on the lessees.

I would reverse the judgment and remand for specific findings on the question of economic coercion.

. The deleted part of the statute provided:

Act of March 3, 1927, ch. 325, 44 Stat. 1365 .

. Act of June 4, 1920, ch. 224, 41 Stat. 759 .

. Act of May 26, 1926, ch. 403, 44 Stat. 658 .

. Irrigable lands under the Big Horn Canal could be leased for ten years.

. In 1947 the Senate Committee on Public Lands reported that of the 575 Crow Indians classified as competent under the 1920 Act, only 318 were still alive; of 2,470 Indians on the Crow Reservation, 2,152 were classified as incompetent. S.Rep.No.368, 80th Cong., 1st Sess. 2 (1947).

. Act of March 3, 1931, ch. 414, 46 Stat. 1495 .

. See note 1, supra.

. See United States v. Labbitt, 334 F.Supp. 665 (D.Mont.1971); Harley v. McCasland, 162 Okl. 117 , 19 P.2d 356 (1933).

. Lessors call our attention to a recent report of the United States Department of the Interior which indicates that 918,412 acres of land are leased by competent Crows while 415,467 acres of land are leased by the Secretary of the Interior acting through the Superintendent of the Crow Agency. According to the lessors, this report found that the average annual rental for the competent leases was $.76 per acre per year while the average annual rental for the leases supervised by the Superintendent of the Crow Agency was $1.65 per acre per year. Effects of Proposed Changes in Crow Competent Leasing Practices, prepared for the Crow Indian Tribe by the Planning Support Group of the Bureau of Indian Affairs (February 22, 1973).