By Carter B. Horsley

The major auction houses have held "intermediate" American paintings auctions each fall at their main auction rooms for many years.

Not this year.

The cutback is surprisingly given the fact that last spring's auctions were not bad. Indeed, they were markedly better than they have been for several years during the depths of the art market's recent depression.

Sotheby's and Christie's did hold American Paintings auctions this fall, but only at their secondary venues, the Sotheby's Arcade, Sept. 30, and Christie's East, Oct. 7. Christie's sale was far superior in its offerings and in its results. Indeed, it illustrated all of its lots in color in its catalogue and Sotheby's had no color illustrations.

The comparisons are as follows:

Sotheby's offered 421 lots, of which 98, or 23 percent, passed, 110, or 26 percent, went for below the pre-sale low estimate, 85, or 20 percent, fell within the pre-sale estimates, and 128, or 30 percent went above the pre-sale high estimate.

Christie's East offered 230 lots, of which 32, or 14 percent, passed, 51, or 22 percent, went for below the pre-sale low estimate, 52, or 23 percent, fell within the pre-sale estimates, and 95, or 41 percent, went above the pre-sale estimate.

There is no rule-of-thumb for what "good" percentages should be at auctions, but in recent years the number of works passed, or "bought-in," that did not sell reached disastrous levels of more than 30 percent and in some sales even higher, enough to scare away many consignors and to petrify the market, to say nothing of individual problems for the sellers confronted with being "burned" and having their assets thoroughly devalued for at least several years.

One always expects surprises at auctions and probably no veteran auction-goer would expect the auction house experts to do much better than 50 percent in their guessing, er, pre-sale estimates. Very often their estimates may not be off the mark, but are not achieved because the handful of anticipated buyers had already exceeded their "allowance," or budget," or because it was a rainy day and the rare "impulse" buyer was absent. Similarly, love is hard to predict and once two collectors with deep pockets set their sights on the same object the results can be astounding.

Generally, of course, the "in-house" experts tend to base their estimates on recent auction sales and are quite conservative. In recent years, however, that conservatism has been pushed extremely far, or, one should say low as the art market's collapse in the early 90's scared the auction houses who, despite their high fees, did not want to get "stuck" with objects that failed to sell and entice any bids. Their psychology of late has been to low-ball estimates and they argue that dedicated collectors are not really influenced by their estimates and that low estimates encourage more potential buyers to get interested. Such an argument, however, usually is self-fulfilling as buyers are influenced by an estimate and generally organize their bidding plans accordingly and also don't like to think they are getting carried away. As these two auctions, and most others, amply demonstrate, the estimates are not scientifically reliable predictors of what an object will sell for.

One must hasten to add, of course, that the in-house experts are under tremendous pressure, both from their business heads and their professional and academic colleagues and peers and they have to deal with an enormous volume of material, much of which needs considerable research, in very short and demanding time-frames. It takes decades for an expert to see enough and such experts rarely stay at auction houses that long before being lured away into the realm of art dealing where their remuneration might begin to better reflect their special knowledge.

Christie's clearly has taken the lead this year in the field of American Paintings and it will be interesting to see what happens at the "important" sales after Thanksgiving.

While the "minor" sales, such as these, rarely have blockbusters, they often offer quite a few gems, amidst a lot of junk, for the beginning collector with a good eye.

At Christie's, there were quite a few good paintings and a very strong market for still-lifes. A very pleasant, very small painting, lot 1, of a tea cup and some bread by John Frederick Peto (1854-1907) went for $29,900. It had been estimated at $7,000 to $10,000. Peto is better known for his pipes and tobacco tins, so the subject matter here was fresh and nicely done.

Much more surprisingly was a larger still life of a cactus by Samuel F. Dubois (1805-1889) that sold for $27,600. It had been estimated at $10,000 to $15,000 although Dubois is exceedingly obscure.

Landscapes, which have been languishing a bit at the auctions, did well. A very dramatic and fine Thomas Doughty, Lot 27, "Fishing on the River," sold for $32,200, well above its much too modest estimate of $12,000 to $18,000. The 22-by-27-inch oil was signed and dated 1828 and is a superb, museum-quality work that rivals the broken-tree-stump wildernesses of Thomas Cole of the same period.

A small scene, Lot 25, of a mountain lake in the Catskills (more likely Lake Mohonk in the Shawangunks) that had been estimated at $3,000 to $5,000 sold for $25,800. Although it bore John F. Kensett's initials, the auction house only attributed the painting as "American School," although the catalogue entry noted that "The sky area of this work was probably painted by John Frederick Kensett." The lovely autumnal scene was reminiscent of Sanford Gifford's work, and was a very fine small Hudson River School painting, whose works have been ridiculously undervalued for the past few seasons.

A few other landscape paintings by quite minor artists such as Alexander Lawrie and Ernest Parton exceeded their high estimates generously.

A delightful 16-by-30-inch painting by William Holbrook Beard of bears shooting hunters, lot 50, sold for $21,850. It had been estimated at $7,000 to $10,000, yet is size and quality deserved more for this eccentric and always amusing artist whose major works are rare.

Other artists whose work exceeded their high estimates at Christie's were Andrew Michael Daesburg, Frederick Judd Waugh, and Isaac Soyer.

At Sotheby's, the stars were Mauritz Frederick Hendrick de Haas (1832-1895) whose "Twilight Over the Harlem River Footbridge," Lot 26, fetched $14,375 and had been estimated at $6,000 to $8,000, and Seymour J. Guy, whose pleasant little painting of little girls, "The Quarrel," Lot 46, brought $12,075 and had been estimated at only $3,000 to $5,000, demonstrating that fine genre paintings still can be appreciated.

Some of the Sotheby's surprises included Edward Wilbur Dean Hamilton's "Lakeside Reflections," Lot 224, which sold for $25,300 and had been estimated at $2,000 to $3,000.

Similarly, Birger Sandzen's "Autumn Glory: Rocky Mountain National Park, Colorado," Lot 240, sold for $20,700 and had been estimated at $8,000 to $12,000. And Rosalind Bengelsdorf's Mural Study," Lot 358, got $12,560 and had been estimated at $3,000 to $5,000.

Hamilton (1862-1943), Sandzen (1871-1954) and Bengelsdorf (1916-1979) are unfamiliar names at American Painting auctions in New York but in recent years a dearth of works by big established names has sent experts scurrying about.