UHLENHOPP, Justice.
These six consolidated property tax appeals involve valuations for 1971 and 1972 of three terminal grain elevators in Sioux City, Iowa. Ordinarily land and buildings are valued as a unit, Maytag Co. v. Partridge, 210 N.W.2d 584 (Iowa), but due to the particular circumstances of these cases the value of the land itself is not. involved. The appeals concern only the values of the elevators themselves, and our holding relates only to the elevators. The elevators have rail and truck but not barge service; they are thus called inland elevators.
I. The applicable legal principles are set forth in our statute and decisions. Code 1971, § 441.21 (all references are to that Code); Tiffany v. County Board of Review in and for Greene County, 188 N.W.2d 343 (Iowa); Juhl v. Greene County Board of Review, 188 N.W.2d 351 (Iowa); and Maytag Co. v. Partridge, supra. The ultimate question in tax valuation cases is the exchange value of the property as a unit — what it would bring between a willing buyer and willing seller if offered for sale. The exchange value may thus be equated to the market value. The exchange value is ordinarily to be ascertained by the sales prices approach. If and only if the exchange value cannot thus be readily established, it is to be ascertained by other factors —what the property would probably bring if offered for sale by consideration of its productive and earning capacity, industrial conditions, cost, physical and function[88]*88al depreciation and obsolescence, replacement cost, and other relevant factors. Even under the sales prices approach, the nature and condition of the subject property must of course be considered in order to compare it to other properties which have been sold and to make adjustments to the amounts of sales prices of such other properties.
II. The present case involves among other problems the question of the burden of proof. We dealt with that question in the Tiffany, Juhl, and Maytag cases. “Burden of proof” may refer to the burden of going forward with evidence or the burden of persuading the fact finder. 9 Wigmore, Evidence, §§ 2485, 2487 (3d ed.); McCormick, Evidence, §§ 336, 338 (2d ed.). Section 441.21 of the Code contains a special provision on the subject; if the taxpayer introduces competent evidence by two disinterested witnesses that the exchange value of the property is less than determined by the assessor, the burden of proof is on the assessor.
Three issues in these cases present burden of proof problems. We hold that each of these issues involves the burden of proof in the sense of burden of persuasion.
The taxpayers employed the sales prices approach as the basis of their cases. To show what these elevators would probably bring if sold, the taxpayers introduced in evidence (1) the sales prices of several other elevators claimed to be comparable and (2) the opinions of two experts predicated upon those sales prices. The first issue, therefore, is whether those other elevators are comparable. The taxpayers have the burden of persuasion on that issue, and it remains with them throughout the case. If they persuade the fact finder — initially the trial court and now this court — by a preponderance of the evidence that the other elevators are comparable, then the fact finder may consider the sales prices of those other elevators and the experts’ opinions predicated on those sales prices, in determining the exchange values of the subject elevators. But if the taxpayers do not so persuade the fact finder as to comparability, then the fact finder cannot consider the sales prices of those other elevators or the experts’ opinions predicated on them. As we will subsequently find, the taxpayers did establish that most of the other elevators are comparable, so that the sales prices of those comparable elevators and the experts’ opinions may be considered.
The second issue, and the ultimate question in the cases, relates to the exchange values of the subject elevators. Taxpayers initially have the burden of persuasion by a preponderance of the evidence on this issue, and they retain that burden throughout the case unless they meet the requirements of § 441.21 of (1) “competent evidence” (2) by at least “two disinterested witnesses” (3) of exchange values less than those “determined by the assessor.” The taxpayers introduced evidence by “two disinterested witnesses” who were well qualified that the exchange values are less than “determined by the assessor,” on the basis of sales prices of other elevators. The question then is whether the evidence of those two witnesses is “competent.” This question turns on whether the other elevators are comparable. If those elevators are comparable, the two witnesses’ opinions constitute “competent evidence” and the burden of persuasion on exchange values shifts to the assessor; otherwise it does not shift. As stated, we will subsequently find that most of the other elevators are comparable. Therefore the two witnesses’ evidence is competent and the burden of persuasion, by a preponderance of the evidence, shifts to the assessor as to the exchange values of the subject elevators.
As to the third issue, in previously applying § 441.21 we have held that the sales prices approach is initially to be used and that the other factors approach may be employed “if and only if” exchange value cannot thus be readily established. Juhl v. Greene County Board of Review, 188 N.W.2d 351, 353 (Iowa). Thus a party relying on the other factors approach has the burden of persuading the fact finder that exchange value cannot be readily established by the sales prices approach. In the [89]*89present cases the assessor arrives at his exchange values of the subject elevators by use of the other factors approach — principally by evidence of replacement cost less depreciation and obsolescence. To make that evidence competent, the assessor has the burden of persuasion by a preponderance of the evidence that exchange values cannot be readily established by the sales prices approach. If he sustains that burden, then the sales prices approach would not apply, and the assessor’s evidence on replacement cost less depreciation and obsolescence would be competent and could be considered by the fact finder in determining the exchange values of the subject elevators; otherwise the sales prices approach would apply. In the latter event the evidence of replacement cost less depreciation and obsolescence could be considered only for any bearing the fact finder deems it has in comparing the physical and functional characteristics of the subject elevators with the elevators which were in fact sold, in the process of making adjustments to the sales prices of those elevators. As we will subsequently find, the assessor failed to establish that exchange values cannot be readily established under the sales prices approach. Of course, although the assessor did so fail, he can still argue for the highest exchange values which can be found under the sales prices evidence and approach.
The trial court initially had the task of deciding these issues of the parties in accordance with the law and the preponderance of the evidence. The court found the issues in favor of the tax officials and upheld application of the assessor’s other factors approach, but reduced the assessor’s valuations by 25%. The taxpayers appealed.
We review de novo. Northwestern Mut. L. Ins. Co. v. Board of Review of City and County of Dubuque, 225 N.W.2d 317 (Iowa). Hence we turn first to the evidence.
III. Grain elevators are of various kinds but we need consider only two: country elevators and terminal elevators. Country elevators are the collection points of grain from producers. Terminal elevators are special purpose properties; they constitute a distinct industry. Most grain moves beyond country elevators by rail, some by truck. Terminal elevators are located at railroad freight-rate break points, where freight rates change. Formerly country elevators extensively used the “transit privilege” offered by railroads, whereby grain stopped at a terminal elevator for checks of weight and grade and then went further on the same freight charge. In those days regulations required that grain moving in interstate shipment had to be stopped for a check of weight and grade at a terminal elevator. Country elevator grain thus largely moved to terminal elevators for direct sale or for check of weight and grade en route for sale to a processor or exporter. Terminal elevators loomed large as an industry, and they actively competed with each other for sales of grain by country elevators and for purchases of grain by processors and exporters. The whole terminal elevator industry was interlocked by the undifferentiated nature of grain and by freight rates, with the prices paid for grain reduced by different freight charges as the terminal elevator and its territory were farther removed from the major grain processors and exporters — the reduction ultimately falling, of course, upon the producer.
Subsequently, the requirement of weight and grade inspections at terminal elevators was rescinded; country elevators could ship directly to processors and exporters. In addition, railroads established multi-car rates which caused some country elevators to bypass terminal elevators. Improved barge traffic also gave an advantage to terminal elevators on navigable rivers; this hurt inland elevators such as the present ones.
During the same period another major change occurred. Until about 1960 governmental policy fostered production of grain and storage of the surplus. Commodity Credit Corporation stored much grain in terminal elevators, at rates favorable to the elevators. Governmental policy later [90]*90changed and brought about reduction of production with corresponding reduction in storage; and Commodity Credit also reduced grain storage rates.
The result of these and other factors shown in the record was that sales prices of terminal elevator properties went down. By 1971 and 1972 a willing buyer would not pay anything like the cost of replacement less depreciation. The terminal elevators in Sioux City were especially hard hit. One factor among several was that South Dakota producers could ship directly to the west coast instead of through terminal elevators in Sioux City.
IV. Since terminal elevators are in a highly competitive industry, scrambling for purchases and sales of grain, the sales of terminal elevator properties are not isolated events; they are interrelated in price. The sales prices of elevator properties can thus be appraised on the basis of cents-per-bushel times capacity of the particular elevator. One witness testified:
The terminal elevators themselves are essentially comparable with one another, since they are all in a vast grain producing area, which this entire Midwest encompasses. They are, therefore, competing with one another, and there are rele-vants between various grain terminal elevators.
They essentially create a market unto themselves, and basically they ship to the same users and consumption points.
They also are affected by the changes of marketing grain in recent years that have been bypassing terminal elevators to varying degrees; but all have been affected in very similar fashion.
All have been affected by the freight-rate structure in grain producing areas to the point of consumption, and all, also, have been affected by the overall decrease in government-owned and storage grain on general economic levels. . . .
They are all in competition with one another to one degree or another, some more than others; but the competing lines within the vast grain area do definitely overlap. They also have other things that are basically common to all of them.
They usually all have to ship to the same ultimate user, whether to the West Coast, or Chicago, or to Texas. They are all competing for a certain amount of business within a geographic area.
Since grain itself is an undifferentiated product, the ultimate consumer is interested in buying the product at the least cost to him; and he does not care whether or not it comes from Sioux City, Omaha, or Minneapolis. It is essentially the same product.
In his testimony the president of one of the present three terminal elevators placed a value on his elevator and stated that on the basis of sales of various elevators,
I am convinced that if we wanted to sell our elevator on strictly an arm’s length transaction — free buyer-free seller arrangement — that is probably what we would get; that our markets are so interlocked, so competitive, a buyer ■ looking for a grain elevator would not pay us $700,000, when he can buy essentially the same elevator for $300,000, even though not in Sioux City. Being in Sioux City is not that big a deal.
Another witness testified:
I think a willing buyer would look at all he’s going to need, when he wants to buy a terminal. He’s going to look at all possibilities. Whether it’s one in Omaha, Council Bluffs, Sioux City, Minneapolis, they all serve essentially the same purpose. And he would look at all those and take the one he wanted. And the business is such that they are all essentially the same. Some are in better condition and some are not.
V. Most terminal grain storage is vertical, but some is horizontal — “flat” storage. In the heyday of Commodity Credit grain storage, some elevators expanded their storage capacity by construction of flat storage buildings. Flat storage presents problems of moving grain in and out, involving additional time and labor. Thus such facilities are best adapted to long-time storage. At [91]*91the same time, however, flat storage presents problems of aeration and temperature control, and grain may go out of condition, especially the bottom several inches. Yet the elevator ordinarily must stand behind the grade — necessitating replacement by proper grain or by money, if grain loses grade.
One expert witness, citing other sales, gave his opinion on the per-bushel exchange value of flat storage considerably lower than for vertical storage. A grain trade witness when asked, “How much would you pay for flat storage?” answered, “I don’t think I would buy a flat storage under the present conditions.” Another witness stated that “it’s more difficult to get the grain into the flat storage than it is into the round bin; then it is piled up on a concrete slab in large piles; that is harder to control the moisture content, harder to control the aeration, harder to control the temperature; and it is not economical to move grain in and out of flat storage. I know of no flat storages anywhere that are being used for merchandising grain, grain coming in and out and being transferred on a frequent basis.”
The retreat of Commodity Credit from large grain storage left some terminal elevators with such undesirable flat storage.
VI. The Bartlett property involved in these appeals is an inland terminal elevator located on United States Highway 75 in Sioux City, Iowa. The site is improved with 11 buildings built from 1915 to 1963, consisting of a grain elevator and workhouse, a large flat grain storage building, a train shed, a truck dump, a grain drier, an office, and a shop. It has upright grain storage capacity of 1,325,000 bushels and flat capacity of 2,751,000 bushels.
The Farmers Union property is an inland terminal elevator in north Sioux City. The site is improved with eight buildings built from 1919 to 1958 with a shed added in 1970, consisting of a grain elevator and workhouse, a train dump, a manual truck dump, a drier, an office, and a shop. It has upright grain capacity of 848,552 bushels.
The Terminal Grain property is an inland terminal elevator located on United States Highway 75 in Sioux City. The site is improved with ten buildings built from 1919 to 1966, consisting of a grain elevator and workhouse, a flat grain storage building, a train dump, a truck dump, a drier, an office, and a shop. It has upright grain capacity of 2,789,000 bushels and flat capacity of 996,000 bushels.
All three elevators suffer from substantial depreciation and obsolescence.
VII. Taxpayers introduced the testimony of two appraisers with extensive qualifications, among other witnesses. The first of these two was George L. Patchin, Jr., who thoroughly studied the subject elevators as well as other terminal elevators. He also studied a number of sales of terminal elevators, inland except as stated, and brought down the list to sales of elevators he considered the most comparable. We initially reject one of them — Archer Daniels to Dixie Portland — as involving special considerations: a large, valuable acreage and a projected flour plant. (We also reject another sale which Patchin stated he gave no consideration; it does not appear to us to be free from compulsion.) Patchin could find no sales of comparable properties in the Sioux City locality itself.
The first sale which we consider from among those described by Patchin was of the Peavy terminal elevator in Council Bluffs, Iowa. It has a capacity of 726,000 bushels, is in very good condition, and has barge facilities. Since it is in better condition and has a river location, it has substantially more value than the subject elevators. Peavy offered it for sale at $650,000 and Cargill bought it in 1971 for $500,000 or 69<t per bushel.
The second sale was of the Fort Dodge Grain Terminal located in Waterloo, Iowa. It has a capacity of 1,000,000 bushels and is a considerably superior facility to the present ones. A five-year lease to Pillsbury granted that company an option to purchase for $180,000 with credit of $12,000 per year on the purchase price, out of the annual cash rental of $21,000. Pillsbury exer[92]*92cised the option in 1971, obtaining a credit of $60,000 under the lease and paying $120,-000 in addition. If the price is considered to be $180,000, the per-bushel price would be 18$; if the price is considered $120,000, the per-bushel price would be 12$. This sale is of importance here, but the facility is materially superior to the subject elevators.
The third sale was for $475,000 in 1968 by Colorado Mill & Elevator to Con Agra of a flour mill and an elevator with a capacity of 2,000,000 bushels located in Omaha, Nebraska. Con Agra considered half of the price to be allocable to the elevator, or $237,500, and Patchin, upon examination of the physical plant, agreed. The elevator is in very good condition, superior to the subject elevators, and at $237,500 would have a per-bushel price of 12.9$.
The fourth sale was for $1,200,000 in 1966 by Lincoln Grain to Archer Daniels Midland of an elevator with a capacity of 8,219,200 bushels — 3,919,200 upright, 2,500,000 flat, and 1,800,000 in round bins. The elevator is located in a suburb of Lincoln, Nebraska. Adjustments must be made for three factors which make this elevator more valuable than average: recent construction of the elevator (1959), excellent condition, and proximity to a bean plant. The elevator has similarity to two of the present elevators in its substantial flat storage. The per bushel price was 14.6$.
The fifth sale was for $162,000 in 1970 by Tonganoxie Grain to Salina Elevator of an elevator with a capacity of 1,895,365 bushels located in Kansas City, Kansas. This property is quite similar in structure and condition to the present elevators. The price was 8.5$ per bushel.
The sixth sale was for $350,000 in 1970 by Anderson Feed to Archer Daniels Midland of an elevator with a capacity of 450,000 bushels located in Emporia, Kansas. The elevator was built in 1960 and is in excellent condition, but steel bins and the absence of a temperature detection system detract from value. The price was 7.8$ per bushel.
The seventh sale was in 1966 by Jennings Grain to Continental Grain of an elevator with a capacity of 4,500,000 bushels located in Hutchinson, Kansas. Jennings advertised the elevator for $1,500,000 and Continental purchased it for $450,000. The elevator was built in 1928 and later, and is in good condition. It has an all-concrete house with convenient hopper bottoms and conveyor tunnels. Its price was 10$ per bushel.
The eighth sale was for $70,000 in 1969 by ADM Grain to Victoria Elevator of an elevator with a capacity of 600,000 bushels located in Minneapolis, Minnesota. The elevator was built in stages beginning in 1906. It brought 11.67$ per bushel.
The ninth sale was for $180,000 in 1968 by Archer Daniels to Burdick Grain of an elevator with a capacity of 2,200,000 bushels located in St. Louis Park, Minnesota. This elevator is quite similar in structure and condition to the present elevators. It brought 8.2$ per bushel.
Based on these sales, Patchin arrived at exchange values for the subject elevators much below the assessor’s valuations predicated on replacement cost less depreciation and obsolescence under the other factors approach.
After Patchin studied the other sales and arrived at his opinions on value, he discovered two further sales in the Midwest grain area. In 1972 Ralston Purina bought Searle Grain’s elevator in Minneapolis, Minnesota, capacity 2,000,000 bushels, very good condition, for $815,000 or 9.2$ per bushel. In 1973 Bartlett bought Missouri Pacific’s elevator in St. Joseph, Missouri, capacity 2,500,000 bushels, very good condition, for $125,000 or 5$ per bushel.
The taxpayers’ other appraiser, Fred A. Herrington, also well qualified, studied some of the same sales Patchin described. His opinions on exchange values of the subject elevators were also much below the assessor’s valuations. The per-bushel range of the values by Patchin and Herrington was 4.4$ to 4.9$ for the Bartlett elevator, 7$ to 8.2$ for the Farmers Union Elevator, and 6.7$ to 10$ for the Terminal Grain elevator. Both Patchin and Herrington ascribed more exchange value to Terminal Grain’s eleva[93]*93tor in 1972 than in 1971, largely due to addition of pollution control devices which were added as a result of litigation.
VIII. The tax officials contend that the elevator sales described by Patchin and Herrington cannot be considered because those elevators are at distances from Sioux City, none of them except Lincoln Grain have flat storage, and because, individually, the elevators possess characteristics different from those of the subject elevators.
We reject the distance argument in these cases largely on the rationale enunciated by the Indiana Supreme Court in Gradison v. State, 260 Ind. 688, 702-703, 300 N.E.2d 67, 78:
The more commonplace the property, the easier it is to find two or more alike or nearly alike. Accordingly, a greater degree of similarity should be required, and the size of the geographic area encompassing the properties to be compared may properly be more restrictive, as an element of similarity. As the property under consideration becomes less commonplace, however, the area may properly be enlarged, without losing what would logically be regarded as similarity. By way of example, the value of a five room single family dwelling house situated in a middle income suburban neighborhood of New York City may be identical to one situated in rural Indiana, but the disparity of other factors relevant to the housing market of the two areas and the availability of other like or nearly alike houses in the same communities render a comparison between the two, for valuation purposes, wholly illogical. On the other hand, accepting gold mines as rare, a valuation comparison, reasonable under the circumstances, might be made between mines, even of different sizes and productivity, although one be located in Colorado and the other in California. Practical considerations enter into such judgments that preclude establishing fixed rules and formulas.
Applying this rationale, we find from the evidence that the other terminal elevators are not incomparable because they are at some distances in the Midwest grain area from Sioux City. See also City of Chicago v. Harbecke, 409 Ill. 425, 100 N.E.2d 616.
As to incomparability predicated on some flat storage in the Bartlett and Terminal Grain elevators, two elements are involved: the effect of the flat storage as to valuation, and as to comparability. We have considered the flat storage valuation problem. As to comparability, we recall this statement in Perry v. Iowa State Highway Comm’n, 180 N.W.2d 417, 420 (Iowa):
Defendant’s objection is that Richards was permitted to tell about the three sales without sufficient foundation and without adequate demonstration of comparability. .
The governing rule is thus stated in Redfield v. Iowa State Highway Commission, 251 Iowa 332, 341-342, 99 N.W.2d 413, 418-419:
“For the evidence to be admitted it must be shown that the conditions are» similar. ‘Similar does not mean identical, but having a resemblance; and property may be similar in the sense in which the word is here used though each possesses points of difference.’ Forest Preserve Dist. of Cook County v. Lehmann Estate, Inc., 388 Ill. 416, 58 N.E.2d 538, 544.”
We note that we made this statement in an eminent domain case tried by jury. An assessor is in an even better position than a jury to weigh “points of difference” and make adjustments accordingly. He has more time, facilities, experience, and expertise in the valuation process.
Adjustments must of course be made for points of difference, but the evidence here shows that flat storage brings a reduction in value and the approximate amount of that reduction. In addition, as differences increase the weight to be given to the sale price of the other property must of course be correspondingly reduced. We find that under the evidence adduced here, the other elevators are not incomparable because of flat storage in two of the present elevators.
The tax officials argue numerous other individual characteristics of the other eleva[94]*94tors which they contend render consideration of the sales of those properties impermissible. The record is very lengthy, and we cannot set out all the evidence. We have examined the evidence on the characteristics of the subject elevators and of each of the other elevators. We agree with the tax officials that we should not consider the Peavy elevator and its sale price because of the barge facilities and because of insufficient evidence to enable us to translate that difference into dollars of value; this is not the kind of a difference for which a court can make an adjustment without evidence. As to the remaining other elevators, we find the differences insufficient to render the elevators incomparable.
We hold that a preponderance of the evidence shows the other elevators, except for Peavy, to be sufficiently comparable to permit consideration of the sales prices of them.
; IX. The tax officials also contend that the sales of other elevators should not be considered and the other factors approach should be employed because under § 441.21 the assessor may utilize the other factors approach when exchange value cannot be “readily established” by the sales prices approach. They argue that the sales of the other elevator properties occurred at points distant from Sioux City, and that they would have to travel thence to examine the properties and would have to engage in extensive computations as Patchin and Her-rington did. Hence, they assert, exchange values of the subject elevators cannot be readily established by the sales prices approach.
We think that “readily established,” just as “comparability,” is relative. Valuing a two-bedroom bungalow in Sioux City is one thing and valuing a large terminal elevator in Sioux City is another, and an assessor must set his sights accordingly. The assessor and his appraiser looked for sales of terminal elevators in the county, found none, and went no further with that approach:
Q. [Elevators’ attorney] Did you use the market value approach to the value in arriving at the figure that you gave to the taxing authorities late in 1971? A. [Assessor’s appraiser] No. The reason for it, of course, was locally in this county I do not know of any sales that have occurred.
Q. Now, if there were no similar properties, similar sales, excuse me, in this district or locale, then you would attempt to apply your replacement cost approach? A. Yes. . . .
Q. [Tax officials’ attorney] Had there been any recent sales of this particular property itself in the recent past? A. Not to my knowledge. There are no comparable sales here in Sioux City.
Q. Were you able to ascertain the fair market value of the property in question through the [sales prices] approach? A. No, because there are no comparable sales in Sioux City. .
Q. [Elevators’ attorney] Moving to the willing buyer-willing seller approach, which you, I believe, said you did not use here because you did not find them comparable, did you confine your search for comparables to the taxing district? A. Yes, sir.
Section 441.21 does not expressly restrict the search for sales of other properties to the taxing district. When from the nature of the property the market for the purchase and sale encompasses a wider area, the wider area becomes the field for investigation. This is necessary in order to give meaning to the sales prices approach in the statute. If in that situation the assessor restricts himself to his county or district, he casts upon taxpayers the burden of performing his task; they must go into the wider field to find sales of comparable properties. If the assessor cannot perform the task with his own staff, he can utilize an appraiser to assemble information.
A case in which exchange value very plainly could not be readily established by the sales prices approach is Maytag Co. v. Partridge, 210 N.W.2d 584 (Iowa). Due to the nature of the property involved here [95]*95and the existence of sales of comparable properties, we hold that the present case is not one in which exchange value cannot be readily established by the sales prices approach.
The result is that the sales prices approach employed by the taxpayers governs here.
X. We arrive finally at the ultimate determination, the exchange values of these elevators in 1971 and in 1972 under the sales prices approach, as shown by a preponderance of evidence. As to the Bartlett elevator, we appreciate that Sioux City, where Bartlett (and the other two elevators) is located, was among the harder hit points in the general decline of the terminal elevator industry; that Bartlett (and the other two elevators) suffers from substantial depreciation and obsolescence; and that Bartlett has a major proportion of undesirable flat storage. Nonetheless, upon consideration of all the pertinent evidence we find that the exchange value of the Bartlett elevator is somewhat higher than the range of 4.4$ to 4.9$ per bushel estimated by the taxpayers’ witnesses. We find the exchange value of the Farmers Union elevator to be toward the high end of the witnesses’ range of estimates from 7$ to 8.2$ per bushel. We find the exchange value of Terminal Grain’s elevator to be somewhat below the upper end of the estimates from 6.7$ to 10$ in view of the flat storage. We thus determine the exchange values of the elevators only, as of January 1, 1971, and January 1, 1972, to be:
1971 1972
Bartlett $207,000 $207,000
Farmers Union $ 69,000 $ 69,000
Terminal Grain $272,800 $312,800
We return the cases to district court for entry of decrees accordingly.
REVERSED AND REMANDED.
All Justices concur except MASON, J., who dissents, and RAWLINGS and REES, JJ., who take no part.