Dream of the Rarebit Fiend: 18 Panels of One Bucking Car

Between panels 1 and 2 of this drawing, the car’s rear wheel leaves the street and climbs about 2.3 cm. The building behind it stays where it was: its sill and street lines sit at the same height in both panels. That is how the whole strip works: within each row the street barely shifts, and the car bucks against it like a horse.

The drawing is an original pen-and-ink board for a Dream of the Rarebit Fiend newspaper strip. The National Gallery of Art holds it as object 142155, accession 2008.128.1, titles it Bucking Automobile, dates it 1907, and has released the image under CC0. The name signed in the lower right corner is SILAS. A Library of Congress catalog note calls Silas a pseudonym of Winsor McCay, and the museum lists McCay as the artist.

What happens in the 18 panels?

The board holds 18 panels in three rows of six, under a title band with a drawn figure at each end. Every panel has a single balloon, and all of them seem to belong to one speaker, a driver in a long coat. In panel 1 he sits in an open car at the curb, complaining that he is “GETTING MIGHTY TIRED OF SITTING OUT HERE WAITING FOR THAT BUNCH!”

Then the car starts to act on its own. It tips onto its front wheels, then its rear wheels, and by panel 4 it stands on end. The driver calls it “A REGULAR BUCK-” and finishes the word in panel 5: “-ING AUTOMOBILE!”

The middle row gets worse. The tires blow out in panel 7, and in panel 8 the car is a squashed ring of flat tires. In panel 10 the car throws him off, and his balloon compares it to a wild horse.

In the bottom row the car lands upside down, and he tries to “TURN HER OVER AND EXAMINE EACH PART.” In panel 17 it bursts into a ring of loose wheels while he cries, “OH! OH! I’M DYING!”

Panel 18 puts the car back at the curb, whole, as three men in top hats come out of a doorway. His last balloon begins “AT LAST, THEY’RE COMING!” and ends “THAT CONFOUNDED CHEESE SANDWICH FIXED ME!” Taken with the title, the bucking reads as his dream.

How far does the car jump between panels?

Open the diagram full size

The jump can be measured. The scan on Wikimedia Commons is 4,000 pixels wide, and the museum gives the sheet as 44.13 × 66.36 cm. That works out to about 60.6 pixels per centimeter, and the scales across and down the sheet agree to within about 1 percent.

In panel 1, the hub of the rear wheel sits 614 pixels below the top of the panel. In panel 2 it sits 473 pixels down, so it has risen 141 pixels, about 2.3 cm on the board. That is about a fifth of the panel’s height, in a single step.

Sideways, the rear hub moves only 2 pixels. The front hub ends up 6 pixels lower and 29 pixels to the right, about half a centimeter. So the car pivots on its front wheel, and nothing is drawn between the two positions.

The reader is left to supply that missing motion. In drawn animation, an animator would usually add drawings for the positions in between, as the Bouncing Ball Animation page shows with frame-by-frame spacing. Here McCay drew only the two ends of the move.

The street holds still while the car bucks

Open the diagram full size

The facade behind the car is drawn again in every panel. It is not a copy, and its details shift from panel to panel: a doorway with a peaked top shows up in panel 3, for one. Its horizontal lines, though, stay at nearly the same height.

In this site’s measurements, the window-sill line could be traced in 16 of the 18 panels. In panels 7 and 10, the spinning blur and the thrown driver cover the edges of the panel where the line is measured. Within each row of six, the sill line sits at the same height to within 5 pixels, about 0.08 cm. So within each row the point of view holds still, and the big movements belong to the car and its driver.

The car, by contrast, goes almost everywhere. The table gives the approximate center of the box this site marked around the car in each panel, as a percentage of the panel’s width (x) and height (y).

Panel 18 puts the car back about where panel 1 had it. The two centers differ by about 1 percent of the panel’s width, less than the marking error: the boxes are marked on the scan to about 15 pixels, roughly 0.25 cm or 2 percent of a panel’s width. So the table is a map of where the car goes, not a precise track.

Download the panel measurements and balloon text (CSV) · source and license notes

Reading the strip panel to panel

This board is a plain grid. Unlike the 1906 Little Nemo board, it has no numbers on its panels. Its three even rows of six point to the usual order: left to right, then down.

The words do something the pictures do not. Four times, by this site’s count, a sentence starts in one panel and ends in the next. Panel 4 ends “A REGULAR BUCK-” and panel 5 begins “-ING AUTOMOBILE!”

Panel 11 ends “WHY THIS MACHINE SHOULD-” and panel 12 begins “ACT THIS WAY, I CAN’T SEE.” The other two breaks fall between panels 13 and 14 and between panels 15 and 16, where “I’LL FIND OUT IF I-” runs on into “DIE IN THE ATTEMPT!”

Each break splits a thought across a gutter, the blank strip between two panels, so the sentence finishes in the next picture. With one speaker and one balloon per panel, the balloons carry the strip’s only voice. For how a balloon’s tail points to its speaker, see Speech Bubbles.

Who was Silas?

The signature reads “SILAS” and is followed by “THANKS TO LONEY HASKELL.” The board does not say what Haskell contributed, and this page does not guess. Below it, a handwritten arrow points up toward the signature, next to a word that seems to read “name.” Farther right, two fainter pencil lines seem to read “7 col” and “Eve Tel.”

McCay’s name comes in from outside the board. The museum’s data on Commons lists the artist as Winsor McCay. A Library of Congress catalog note for another Silas drawing, dated 1906 (LCCN 2005680624), says “Silas was the pseudonym of Winsor McCay.” That note reached this page as it appears on Wikimedia Commons; the Library’s own record was not consulted directly.

The same note reports McCay’s own story of the name. According to the note, he said Silas was the name of an old man who drove a garbage cart outside the New York Herald office. It adds that he used it to keep these strips apart from his strips for children. For this the note credits a biography of McCay by John Canemaker, which was not read for this page.

A trade paper made the same link in print. Moving Picture World, in its issue of November 14, 1914, described “Winsor McCay, creator of ‘Little Nemo,’ ‘Dreams of a Rarebit Fiend’ and other popular newspaper cartoons of international fame.” The paper’s title differs from the board’s by a plural and an article. The notice was about his hand-drawn film Gertie the Dinosaur.

What does the board date?

The year comes from one line along the bottom edge: “COPYRIGHT, 1907. BY THE NEW YORK EVENING TELEGRAM. (NEW YORK HERALD CO.).” It names the newspaper and, in parentheses, the company behind it. The pencil “Eve Tel” in the corner may point the same way, but it is only a pencil note.

The board gives no day. It shows no month, day or issue number, and the newspaper page it was drawn for was not read for this page. So the day this strip ran in print is left open here, and the board is dated by its year alone.

The museum describes the medium as pen and black ink and brush and black ink over graphite, “with printed overlay,” on paperboard. The record does not say which part is the printed overlay, and the board itself does not settle it.