Rankin, C.J:— The defendant appeals from the decree made in a suit brought for purpose of setting aside the sale of a holding in Panchannagram. The sale was held on 21st December 1926, for the default of payment of revenue or rent, which accrued due on 28th March 1926 and for which the latest date for payment, under the orders of the Board of Revenue was 28th July 1926. This is described as the rent or revenue for the Bengali year 1332, which corresponds to the period 14th April 1925 to 13th April 1926. Apart from a subsidiary question, Which has reference to certain buildings on the land, the only question upon this appeal is the question whether there was any arrear of revenue for 1332 to justify the revenue sale.

2. The terms of the tenancy are to be collected from the kabuliyat executed by the plaintiff's predecessors-in-title and dated 19th April 1876, a few days after the beginning of the Bengali year 1283. The kabuliyat states that the rent is to be at the rate of Rs. 6-6-6 per year and that the tenants will deposit the said rent in the Collectorate within 28th March each year. According to the contract therefore the rent for a year was not payable in advance, but was payable about a fortnight or three weeks before the end of each year of tenancy. The year of tenancy is not stated to be the Bengali year, but if it extended from 19th April of one year to 11th April of the next, it coincided, save for a few days with the Bengali year. The financial year or official year, for purposes of revenue, begins on 1st April and ends on the last day of March. In order to determine whether, in December 1926, the plaintiff had failed to pay the sum of Rs. 5-15-6, the sum to which the rent of Rs. 6-6-6 had been reduced, due in respect of his occupation of the land for the Bengali year 1332, or perhaps more accurately for the period 19th April 1925 to 18th April 1926, which is the question before us, we have to examine the jama wasil bakis of this holding. These are accounts kept by the office of the Collector of the District of 24-Pargannas and the dispute between the parties is as to their meaning. In other words, as the Judicial Committee pointed out in a case of similar character having reference to Dihi Panchannagram: Narendra Nath Dutta v. Abdul Hakim . AIR 1928 PC 243, the question before us is a question of fact: whether a certain sum was paid before a certain date or then remained unpaid, and the entries in these jama wasil bakis are to be regarded as a narrative of monetary transactions, a narrative of which the meaning is to be got at by considering what is said and the manner of saying it.

3. The accounts produced begin with the year 1899 and continue down to the date with which we are concerned. The first thing to be collected from them is that from 1899 down to 1910, the tenant paid a sum equivalent to the annual rent in June of each year. Thereafter down to 1924, he paid his rent in each year generally in July, once or twice in June and once or twice later. In the year 1921, for some reason, he paid twice, paying Rs. 6-6-6 in addition to the annual demand, which at that time was Rs. 6-1-9 for the first part of the period comprised in these accounts, namely, down to 1914, the last date for payment of arrear of rent for the purpose of avoiding a revenue sale was 28th June of each year. Prom 1914, it was 28th July each year. Now the sole question upon these accounts is this: whether they show that the payments, which were made, were payments of rent in advance, in the sense that the payments made in June or July were made in respect of the rent due from the tenant in respect of his occupation of the land from April immediately before the payment was made until the March of the following year, or whether they show that they were made in respect of the year of occupation which ended in the previous April and for the rent which became due under the contract on the 28th of the previous March. This is the same question which has given trouble before in at least two cases in this Court, but on a question of fact I propose to take no notice of other cases and to examine the documents by themselves.

4. It appears to me to be demonstrable on the face of these accounts that the latter of the two alternatives above stated is the correct one. We are not asked to assume, nor could we assume, that any arrear or overpayment existed, prior to the period dealt with in these accounts. As the entry for 1899 is defective, in that it puts no date to define the period in respect of which the payment is being made, it would have to be filled up, if at all, consistently with the subsequent entries. We must either neglect the first item altogether, or we must fill up the gap, as if the “demand” were defined by the year 1899-1900—it makes no difference which we do. Taking the next three entries, the question is what is meant by “demand for 1900-1901,” “demand for 1901-1902” and “demand for 1902-1903.” Does it refer to the rent which became due on the 28th March of the first of the two years mentioned—i.e, in 1900, 1901 and 1902—or does it refer to the rent which at the time of payment in June had not yet become due and would not become due until the following March? We may bear in mind on this question that we are dealing with accounts kept by a revenue officer and that 1900-1901 is the financial or official year. The learned Advocate General for the plaintiff respondent, contends that each year, when the tenant paid his annual rent in June, he was paying rent which, under the contract, would not be due until the following March and for which the tenant had, until the following June, time in which to pay, for the purpose of avoiding a revenue sale. This he contends must be the meaning of the words “current rent” or “current demand.”

5. Now, in the year 1903, the accounts as kept show a change of method. The date 1902-1903 is followed by “1309 B.S” and in each June, as the tenant pays, he is said, for the next five years, to pay in respect of the demand for 1309, 1310, 1311, 1322 and 1313 respectively. In 1908, the former method is reverted to and the demand is defined by the year 1908-1909. This goes on consistently till 1915-1916. Again a change of method is introduced and the account reintroduces the Bengali year as defining the demand in respect of which the payment has been received 1915-1916 is followed by 1322 B.S and thereafter the account continues in terms of the Bengali year.

6. Now, if we are to extract the meaning of an account the first assumption which we must make is that the account may be consistent and correct. An interpretation, which avoids imputing to it glaring errors, is to be preferred to an interpretation, which involves the conclusion that an error has been made and successive further errors have counterbalanced its effect. We will therefore examine the three transition periods in this account, concentrating upon the fact that the year 1902-1903 is followed by 1309 B.S, that the year 1313 B.S, is followed by 1908-1909 and that the year 1915-1916 is followed by the year 1322 B.S If we are not to assume that the Collector's officer is in each case making a highly serious mistake, it seems to me to follow that the payments made by the tenant in June or July had reference to the rent which was due on the previous 28th March in respect of the year of occupation which concluded on the 18th April thereafter and which was finally demandable for purpose of revenue sale in the middle of the calendar year. From 28th March till 18th April, the rent for the previous year had certainly been due under the contract. That however was a short period of three weeks and during most of these three weeks the year of occupation had itself not come to an end. April 1902, for example, to April 1903 was the Bengali year 1309. The payment made in June 1902 is put down as referable to the demand 1902-1903; payment in June 1903 is defined as the demand for 1309. If, in June 1902, the tenant had paid his rent from 19th April 1902 till 18th April 1903, then when he comes next June to pay his rent in 1903, how can that rent lie put down as referable to 1309 B.S? Clearly it cannot 1902-1903 is the revenue year in which the Collector had to see that he got his rent by the 28th June, for the year of occupation which ended in the middle of the previous April. The year “1309 B.S,” shows that a change is being made to define the demand in a more simple and more natural way by giving the Bengali year for which the rent was paid—that is the year of occupation in respect of which the rent was due. If the original method be continued after the year 1902-1903 until we come again to the year 1908-1909, it will be found that there has been no mistake. 1908-1909 comes in its proper place consistently with 1902-1903.

7. Let us look then at the change from the method employed to record the payment made in June 1907 to the method employed in recording the payment made in June 1908. From 1313 B.S, we change in the following year to 1908-1909. The year 1313 began on 14th April 1906, and ended on 13th April 1907. Why should it then be followed by the figures 1908-1909? Are we to say that when 1902-1903 was followed by 1309, the tenant is recorded as paying rent twice and that when 1313 B.S, is followed by 1908-1909 a year has been allowed to drop out? Clearly not. By the rent due for 1313 B.S, we mean, or any ordinary tenant would mean, by that expression, the rent which he had to pay by June 1907 if he wanted to avoid a revenue sale. The rent for the following year 1314 B.S, that is, April 1907 to April 1908, was demandable on pain of revenue sale by June 1908. It is true that for three days before the beginning of the financial year 1908, it had been payable under the contract, though the period of occupation extended for a few days beyond the end of the Bengali year 1314. The successive entries are quite consistent, if it be remembered that when the year is stated in the Bengali style the year referred to is the year of occupation in respect of which the rent is due and that when the year is stated by double figures the year meant is the financial or official year in which the Collector's duty to recover it arose and in which the Collector could reasonably expect to receive it. The same reasoning applies to the transition from 1915-1916 to 1322 B.S In both years the payment was made in July and in the second it was made at the 11th hour. The learned Subordinate Judge, consistently with the contention before us, was invited to hold that here again the same rent had been paid twice and solemnly recorded by successive entries in the books. He has accepted this contention and says “this appears to be clear.”

8. An examination of the jamawasil baki will show that the change of method which had been introduced in 1903 and abandoned in 1908 was again being introduced, but that it was introduced more carefully with an explicit statement as to what was being done. As the learned Advocate-General's contention involves that the payment made in June 1903 as being for 1309 B.S was a second payment for the same period as had been covered by the payment in the previous year, it may well be that simple-minded tenants had been attracted by the same argument and that the reversion to the old system in 1908 was not unconnected with a certain amount of confusion which the change of 1903 had caused. Be that as it may, we see that after 1916 the Bengali year is deliberately adopted to define the demand in respect of which the payments are received; and we find further that in the transition year the demand is carefully defined.

9. The payment received in July 1916 is stated to be current for 1322 up to 28th July 1916. This can only have reference to the rent which, on 28th March 1916, became due under the contract in respect of the year of occupation which ended on 18th April 1916, and which was finally demandable under the orders of the Board of Revenue on 28th July 1916. This entry, like the entry of 1309, shows one that the previous entry had reference to the financial year which began a fortnight before the end of the period of occupation in respect of which the rent was due. 1322 B.S is not a financial year at all. The change is not merely from what may be called the English style to the Bengali style. The Bengali style tells one that the demand is being more simply defined by a reference to the Bengali year, which in everybody's mind is the year during which the tenant has enjoyed the land and for which the sum is due. The period of occupation could not naturally be defined by the financial year which for purposes of collection and account is of primary interest to the revenue.

10. If this be right, the controversy is really at an end. There was no double payment in respect of the year of occupation 1322. The rent for 1323 was timeously paid in July 1917. In 1324, the rent had dropped from Rs. 6-6-6 to Rs. 6-1-9 and this was timeously paid in July 1918. In July 1919 rent was paid for 1325 at the old figure of Rs. 6-6-6, which was an overpayment of four annas nine pies. Taking this overpayment into account, the rent for 1326 was paid in September 1920. For 1327, there was for some reason a double payment, one of Rs. 6-6-6 and another of Rs. 6-1-9 so that the account was Rs. 6-6-6 in credit. The rents for 1328 to 1330 were at the reduced figure of Rs. 5-15-6 and were duly paid, though in the last case not until December. In 1331 nothing was paid, but the account being Rs. 6-6-6 in credit, there was no arrear, seven annas still standing to the credit of the tenant's account. For 1332 nothing was paid in 1926, and then in December 1926, the holding was sold for arrears of revenue. The Collector was quite within his rights.

11. When the case of Narendra Nath Dutta v. Abdul Hakim was decided in this Court by Chatterjea and Graham, JJ., they came to a conclusion of fact upon jamawasilbakis, in which the entries may have been of the same character as those before us, though the contractual date for payment was 28th June and the year of occupation ended 6th July, dates which make a good deal of difference. I do not propose to discuss the facts of another case upon a question of fact. But it may be worthwhile to notice that the decision in this Court in Narendra's case was given in August 1924 and that on 4th June 1925 the same matter of these holdings in Panchannagram came before Walmsley and B.B Ghose, JJ., in an unreported case Ramlal Chaudhuri v. Bijaygopal Mukherji . 1925 AOD 331 of 1922.. In that case, as in the previous case and in this case, the parties had taken copies of these technical accounts and had proceeded in the trial Court to dispute about their meaning without taking the trouble to call any evidence from the Collectorate by a person familiar with these accounts who might be qualified to explain them. Walmsley and B.B Ghose, JJ., thought fit in this Court to receive the evidence of a clerk from the Collector's office. They appear to have come to the conclusion that the

“demand each year is shown as being for the period ending 28th July, that is to say, in the account for the Bengali year 1325 or the financial year ending 31st March 1919, the demand is shown as being for the year ending 23th July 1919.”

12. I cannot take into account the evidence in another case, even if I had the record of it. Proceeding entirely upon the internal evidence of the accounts before me, which are quite explicit after 1916, I conclude that the accounts are consistent throughout, no years being dropped out and no payments being recorded twice in respect of the same year. During the period, for which the account is kept in terms of the financial year, there is, in this case, only one instance of payment being made after the Collector's final date in June or July. In that case the payment is entered as being in respect of an arrear. But the figures 1913-1914 do not occur in the printed sheet, though the payment was undoubtedly received in that financial year. Upon this account, I think it enough to say that the demand is shown, where the financial year is employed, as being for the financial year in which the rent is finally demandable, that is in which on 28th July the holding becomes liable to sale, and that the demand is shown as “current” up till 28th July of that year. I may here observe that for each payment the tenant gets a chalan or receipt, which presumably shows full particulars of the rent which he has liquidated. The tenant has not produced any one of these, preferring doubtless that his arguments as to double payment should be considered without them. My finding of fact upon the evidence is contrary to that of the learned Judge and I think the appeal should be allowed. It has been called to our attention that the plaintiff has certain buildings which at the date of suit stood on the holding. As the learned Subordinate Judge appears to have expressed an opinion that the plaintiff's right, if the sale were valid, would be to retain possession of them upon the payment of an equitable rent to the defendant we are asked by the appellant to negative this view. While it is clear that at a revenue sale the buildings do not pass with the land [Narayan Das v. Jatindra Nath Roy . AIR 1927 PC 135], the learned Advocate-General has not defended the view that the defaulting tenant is entitled to retain possession of the buildings on the land upon payment of an equitable rent. The appellant is willing that the plaintiff should, if he so desires, remove the buildings; or if he does not desire to remove the buildings, the appellant is willing to take them at a reasonable valuation. This does full justice to the rights of the respondent and we will order accordingly. The result is that the appeal is allowed with costs and the plaintiff's suit is dismissed, with costs. But in our decree we will make the order which I have indicated as regards the buildings.

Mitter, J.:— I agree entirely with my Lord, the Chief Justice both in the reasons and conclusions of his judgment in, this case.

K.S

13. Appeal allowed.