« 'A Degree Can't Hold Your Hand' | Main | MTF Representation »

Sunday, 30 September 2007


TrackBack URL for this entry:

Listed below are links to weblogs that reference Which Came First, 'On' or 'Off'?:


Mike, I think the certitude with which you offer your own analyses and justifications-- pro or con, ON or OFF, expert or layman-- is exactly the point in Morris's blog. Further, your offering of Gardner's sky painting has nothing to do with the question at hand evidenced in Fenton's photographs.

What makes you the expert, any more than anyone else, able to proclaim that a vast number of the respondents are incorrect? Statistically-speaking you're probably correct, but the reality is that you *think* you know the answer. But you don't actually know the answer-- at least not any more definitively than anyone else-- and again I think that's one of the points Morris is trying to make. Proclaiming or even possessing expert insight doesn't get to the truth of the matter any more than does conjecture.

Frankly, I think the ultimate answer (or "truth") to the mystery is the least interesting thing about the essay. I think Morris is brilliant, and I think we'll see him expound on these issues of truth and certitude and how we navigate these issues in regard to imagery in his forthcoming book. In the meantime, I'm really amused and entertained with the certitude and ferventness with which people argue about their beliefs surrounding Fenton's mystery.

Personally, I think there are 2 questions here:
1. did he pose one shot?
2. which came first?

As to the first question, there seem to be an awful lot more cannon balls in the ON picture than OFF. If he posed them, why not just take balls from the ditch to the nearest point on the road or vice versa. this suggests balls wold have been moved to/from a location out of view. Why do that? Seems a pointless waste of effort & there are a lot of balls to move (I count around 30 on the road).

That would lead to the point that OFF would come second only if balls had been harvested but that seems incongruous. First aim would be to clear the road, for sure, but why only take a few, randomly scattered balls from the ditch? Also seems a bit odd.

I'm going to come down on the side of OFF then ON and the fact that a bunch of cannonballs came careering in during the 2h.

You really need to read my post before you comment on it.


Simplistic analysis: in the "On" photo, there are many more balls, both on and off the road. The "Off" photo has cannonballs, but many fewer. Cannonball scavengers. The balls left may have been damaged; scavange the good ones. This was probably common on both sides of the conflict; some of the balls would eventually be damaged.

An aside, I am unsure that there is a real problem. A story is told, maybe more dramatically in "On". In this age of reason, when are we going to get over the notion that photography tells the truth? "Journalistic integrity", "honest politician" Pfui, even when those phrases are accurate, a picture does not always tell the "truth", even if it is accurate, and unaltered.


I don't know whether he faked it or not, but I do know from Civil War stories that cannon balls were often seen as they bounced and came into the lines of soldiers, and that they came fast and hard, much faster and harder than a thrown baseball, and they were very heavy. If a bouncing cannon ball hit you, it could kill you or take your leg off, and because they bounced in crazy ways, you couldn't predict exactly where they were going. Also, these balls were intended to explode, and the ones you see are the ones that didn't. If a cannon ball's fuse was poorly cut, or the fuse wasn't quite right, the ball might bounce and stop and then explode, seconds after it was intended to. (There is actually a scene I've read somewhere of a ball bouncing into a group of waiting cavalry and sitting there hissing, as its bad fuse sputters and finally goes out, as the cavalry guys stand there and sweat.)

In any case, it would seem to be an insanely risky proposition to be picking up and rearranging fresh cannonballs, just for a better photograph. Not to say that it wasn't done, but it seems crazy to me. (Harvesting would be a different matter -- you could assume that the cannon balls were dead, after they'd sat for a while.)

But then...if the the people cited in the original column are correct, and Fenton was shooting back at the British lines, does that mean he drove his wagon over those cannon balls going in? It doesn't seem likely that he'd have set up in a place where the balls were skipping around him to get to the road.



I think you're misreading something, Mike. Wendy Ju and Mark Klett are in agreement that OFF came before ON, though for different reasons--

Klett: "Based on the lighting between the two photos the photo with the cannon balls (ON) was taken AFTER the one without(OFF)."

Ju: "Using entropy as a guideline (things tend to go from whole to broken, from high to low, undented to dented, etc), it seems like it is more likely that the OFF image preceded the ON."

And they both agree with Sontag.

I agree with Klett's approach of looking at the highlights rather than the shadows (and stated so in comment #239), with the caveat that it's hard to tell from the low res images if exposure is playing a trick here. I found the high res LOC image, but without both of them, it's hard to tell anything more.

I like your observation that he could have been shooting under fire, meaning ON came after OFF, but wasn't staged, but I can't imagine why they wouldn't just get out of there if that many cannonballs were coming at them, unless maybe it was a barrage and they took cover, and then when all seemed quiet, they made a photograph before heading home.

By the same token, I don't see them rearranging all those cannonballs by hand in 90 minutes and having enough time to make two wetplate images.

Or.....in the time Fenton and his "assistant" spent at this site, they actually were farther back avoiding the bombardment that littered the area, leaving the camera in position to take it's chances, thus an answer to everything, i.e. #703.

Now I need to go lay down.


P.S. Mike, I enjoyed your analysis, some very interesting technical comments.


William Frassanito in his book, Gettysburg, A Journey In Time, does a thorough analysis of the Gardner photo and pretty well proves that the body was moved to enhance the composition. This wasn't uncommon during that era. It's important to remember that Gardner was a commercial photographer not a photo journalist. There are several photos from Gardner's Gettysburg work that show a rifle that seems identical to the one in the sharpshooter picture suggesting that Gardner used it as a prop throughout the series.

Mr Frasssanito's research shows that Gardner invented stories and deliberately mislabeled photos to fill in gaps in his coverage of the Gettysburg battlefield.

Mr. Frassanito has done extensive work documenting camera locations for Civil War photography. His books are fascinating reads on several levels not the least of which is setting the historical record straight on some of the iconic photos from the 19th century.

All of this leads to the point that while Gardner is not Fenton and it would be perhaps unfair to paint him with the same brush, the photographer narratives surrounding many of the war photos from the mid 1800's have to be taken with a grain of salt. It's entirely possible that Fenton moved cannon balls to enhance his photos.

Randall Teasley

Whoops, sorry, typo. I meant to write that Klett AGREES with Wendy Ju.

I make so many typos in the course of a day it's frankly amazing that I manage to keep this blog going.


While far from definitive, I think a clue that supports the OFF before ON lies in the lower middle cluster of cannon balls. There are a few balls that seem to have nestled themselves quite nicely amongst some of the existing balls ("off" photo) without disturbing their positions. That seems unlikely as cannon balls would be carrying a great deal of momentum.

As well, there are a number of balls on both the left and right sides of the "off" image that are missing in the "on" image. As the "on" photo contains MORE cannon balls in both the ditch and on the road, it's suspicious that there are fewer balls located on the sides of the frame. Common sense would dictate that there should be more balls everywhere.

What I know of forensics I learned from CSI, so forgive me any gaps in my logic.


I like your addendum; I also agree with you, the more I've looked at the images, in spite of my somewhat disdainful feelings about the argument in general, though you have answered that very well. Obviously laying down didn't help.
Maybe I'll go do some late work and get this off my mind.


If you had taken the more dramatic ON picture first, why take the OFF plate 90 minutes later? What does it bring that the ON did not have?

Better sharpness or exposure maybe ;-)

The explanation that something happened that made the scenery more interesting after an hour or so sounds quite appealing...

Dear Gitane,

I'm not sure what you think Mike wrote, but statements like, "As for Errol Morris's main question... I find myself 65% convinced by the excellent analysis by Wendy Ju of Stanford, although I think it's likely that the truth cannot be known to any reasonable degree of certainty" can hardly be considered 'certitude!' I don't think you can qualify a judgement much more than that and still have one!

(Well, not by more than another 15%, anyway [grin].)

Some truths -- Expertise DOES trump voting numbers. That's ultimately the definition of an expert-- someone who is right substantially more often than the average person. Otherwise, everyone would be equally expert. And that means what Mike, as a qualified expert (and would you really disupute that?) thinks *IS* worth more than what some several hundred unqualified people think.

Some matters are a popularity contest. Many more are not. Everyone may be entitled to an opinion (a metaphysical question I'm not mentally equipped to address) but all opinions are most assuredly not created equal.

As for the photographic issue itself, I don't have an opinion. I would observe that there are some interesting changes in the landscape at the left (vis Wendy's blink comparison). Rocks have shifted in seemingly arbitrary ways and it looks like there's a cannonball that's been moved either into or out of a shallow depression. The changes are sufficiently plausible running time either way that I'm disinclined to comment beyond that. But it's clear that no matter which way time runs, the landscape got substantially messed with, either intentionally, accidentally, or by bombardment.

Makes it kinda hard to judge the photog's intent from first principles. It's not a clean 'crime scene.'

pax / Ctein

I also brought the photos into Photoshop...

First, as already noted in the comments of the article, the photos do not align exactly. It is obvious :-) that if Fenton was posing the ON photo, he would have left his camera on tripod while the cannonballs were moved to the road.

Second, the OFF photo has larger dark areas. That is, the cannoballs in the ON photo are lit more. That would say the ON photo was later, when the sun rose (more) in the sky, even if it was cloudy. Well, unless Fenton simply used a longer exposure, although the softness of highlights in the ON photo argues against that. And the background hill is also lit better in the ON photo. It's almost lost in the haze in the OFF photo.

Third and last, there are more cannoballs in the ON photo. For instance, in the lower right foreground on the OFF photo, there's one cannonball. In the On photo, there're five. There's also an additional ball a fingerwidth from the lower edge mid-left, right beside the trail.

So the suggested sequence of events is: Fenton comes to the spot, takes the OFF photo. Moves back. The Russians start shelling the spot some time after that. After the shelling stops, Fenton comes back, positions the camera on almost the same spot and takes the ON photo.

Man, this CSI stuff is fun, even if I'm wrong. :-)

Mike, I came to the same conclusion you relate in your addendum, based on the same passage if not the same sentences. I had to speculate that we are looking toward the target zone from which he had to retreat 100 yards, but your reading makes that irrelevant. (Curiousity about focal length lingers, however.)

I find it implausible that someone would put their life at risk to manually add shot to a target zone during a bombardment, just for a better picture. Of course, those calling Fenton a photographic liar might simply say he is lying as well in his letter, but that double accusation should carry a greater burden of proof.

Another point: I found Professor Klett's analysis confusing. When facing south, right is west, isn't it? Not east as he seems to be saying? Regardless, if the sky in the photo cannot be discerned, as you convincingly explained, then can any analysis of the directionality of ambient light, which can be affected by cloud conditions, be conclusive?

There's a lot of conjecture on both sides. The major thing that puzzles me is that there are no craters. Craters would indicate that in the bombardment some of the shells had exploded. With so many balls on the ground and with the absence of craters it would seem ludicrous that the Russian Artillery had such a high failure rate with shells not exploding. Solid canon ball were usually used primarily against fortified position or ships. Grape and chainshot were used against cavalry on land to cause maximum casualty. From reading accounts of the Charge of the Light Brigade we know that the Russians were highly skilled with the use of artillery.

So my conclusion....well I don't know which came first, but things are not as they seem.

I'm no expert but those look like muzzle loading cannon balls to me which wouldn't be designed to explode.

"I'm no expert but those look like muzzle loading cannon balls to me which wouldn't be designed to explode.

Posted by: Paul Mc Cann"

Here is a quick list of the different types of projectile fired from a canon of that period:

1) Round shot
A solid projectile made, in early times, from dressed stone but, by the 17th century, from iron. The most accurate projectile that could be fired by a smooth-bore cannon, used to batter the wooden hulls of opposing ships, forts, or fixed emplacements, and as a long-range anti-personnel weapon.

2)Chain shot or bar shot
Two sub-calibre round shot (a good deal smaller than the bore of the barrel) linked by a length of chain or a solid bar, and used to slash through the rigging and sails of an enemy ship so that it could no longer manoeuver. It was inaccurate and only used at close range.

3)Canister shot
An anti-personnel projectile which included many small iron round shot or lead musket balls in a metal can, which broke up when fired, scattering the shot throughout the enemy personnel, like a large shotgun.

4)Shrapnel or spherical case shot
An iron anti-personnel projectile containing an interior cavity packed with lead or iron round balls around a small bursting charge of just enough force to break open the thin-walled iron projectile. A powder train in a thin iron sleeve led to a time fuse inserted into a holder at the outer edge or the projectile. The fuse was designed to be ignited by flame from the propellant charge. Ideally the case shot fuse would detonate the central bursting charge when the projectile was six to ten feet above the heads of enemy infantry thereby showering them with the iron balls and fragments of the casing. (Invented 1784 by Lt. Henry Shrapnel, Royal Artillery, Great Britain).

An explosive anti-material and counter-battery projectile, of iron with a cavity packed with a high explosive bursting charge of powder used to destroy enemy wagons, breastworks, or opposing artillery. Two types of fuses were used—impact fuses that detonated the bursting charge by percussion, and time fuse cut to length measured in seconds and ignited by flame from the propellant charge

An anti-personnel weapon, similar to canister shot, but with the shot being contained in a canvas bag, and generally of a larger calibre. So called because of the resemblance of the clustered shot in the bag to a cluster of grapes on the vine. In one variation of this, the shot was held together by a coiled bar, and was spread by a fused charge in the same way as a shell.

An incendiary/antipersonnel projectile designed to burn fiercely and produce poisonous fumes. It was constructed of an iron frame bound with sack cloth and filled with various ingredients such as pitch, antimony, sulphur, saltpeter, tallow and venetian turpentine. It was ignited by the cannon's propellant charge, bursting on impact with the target and releasing noxious fumes while setting fire to its surroundings. It was effectively an early chemical weapon as well as an incendiary and area denial weapon.

So as I said in my previous post the absence of craters means that the Russian gunners were incredibly inept in not getting shells to explode or this collection of ballshot came from somewhere else, and things in both photos are not what they seem.

I'm largely convinced by Wendy Ju's analysis that OFF preceded ON. Mark Klett's method of analysis leads to the same conclusion, but I think he confuses left and right in his writing.

Several factors suggest to me that,Fenton 'staged' most (if not all) of the cannon shot in the ON photo.

First, if all new cannon shot in ON were 'placed' by Russian cannon fire, several indicia strongly suggest that the direction of fire was from behind the camera into the frame. That is, the camera was pointed away from the Russian cannons and toward the edge of the firing range. Klett's analysis suggests the camera is facing southwest. Several of the 'experts' in the Morris article believe Fenton was facing south, and indicate south would be toward the British encampments.

Second, while Fenton wrote that the cannon shot was "coming up the road," OFF and ON are facing 'uphill.' If the Russian cannons were in front of the camera, the cannon shot would have been coming 'down the road.'

Third, the 'distribution' of new cannon shot indicates that the camera was facing the outer edge of the firing range. Fenton states in his letter that he did not shoot at his desired location because it was in the line of fire, but instead retreated to a location about 100 yards short of his desired location. At that location, a cannon ball rolled to his feet. This suggests that he retreated to the edge of the firing range. If Fenton set up his camera at that location to face the cannons while the Russians are firing those cannons, the distribution of new cannon shot in ON should increase with distance from the camera. By contrast, if Fenton set up the camera to face away from the cannons, and toward the edge of the firing range, the number of new cannon shot in ON will decrease with distance from the camera. A review of ON indicates the latter scenario: all of the new cannon shot are in the lower half of the frame. Assuming all new cannon shot are from Russian cannon fire, the photos indicate that camera was facing away from the cannons to the edge of the firing range. Which means that all new cannon shot in the photo went past the camera location.

So, either Fenton set up his camera within range of 'slowly rolling' Russian cannon fire, with the cannons behind the camera, and the cannon fire rolled shot past the camera location to create the scene of ON; or Fenton moved the cannon shot onto the road and into the ditch to create the scene of ON.

I strongly tend to the latter conclusion for the following reasons. First, Fenton suggests, in his letter, that one ball rolled to his feet and plenty of cannon shot continued to pass by on each side, but not near to him. While I cannot definitively state what Fenton believed was 'near,' I would describe the new cannon shot in the foreground of ON to be near to the camera location.

Second, the distribution of new cannon shot in ON appears inconsistent with physics. Cannon shot should tend to collect in 'low places.' An analogy to bowling: Once a ball is in the gutter, it tends to stay in the gutter. Indeed, it would require significant force directed at an angle to the gutter for the ball to leave the gutter. Since Fenton ended up at a location of 'slow rolling' cannon shot, they would tend to lack sufficient potential energy to leave the ditch once in the ditch. As demonstrated in OFF, ditches (rather than roads) would tend to collect cannon shot.

Third, the 'angular' size of the road from the perspective of the Russian cannons would indicate a decreasing probability of cannon shot rolling to a stop on the road. A further analogy to bowling: the longer the alley, the more likely the ball will end up in the gutter, especially if the bowling lane isn't straight and level. Since this is at the outer edge of the Russian cannon range, the distribution of 'resting places' for cannon shot would have increased the tendency of cannon shot to end up in the ditch.

A quick count of new cannon shot in ON reveals, however, that twice as many new shot are on the road rather than in the gutter, er, ditch.

Fourth, the new cannon shot in ON appear in two fairly distinct groupings: on the road and in the ditch. On the road, the distance of the new cannon shot from the camera extends to nearly 1/2 way up the photo. In the ditch, the distance from the new cannon shot to the camera extends about 1/3 way up the photo. Interestingly, beyond this 1/3 point, the ditch had collected a significant number of old cannon shot (that is, cannon shot seen in OFF).

I would suggest that this is an impossible distribution of cannon shot for distant cannons to achieve, but it makes sense from a 'composition' viewpoint. Adding 'shot' further up the ditch would have little (or no) impact. The further portions of the ditch already have a significant number of cannon shot, although their angular size at that distance from the camera is quite small. As such, any additional cannon shot further from the camera in the ditch would barely be visible and have little compositional impact.

By contrast, adding shot to the foreground ditch and onto the road has obvious impact.

Fifth, and speaking of 'impossible distributions' of new cannon shot, I note that six new cannon shot are located in ON at the the very front edge of the foreground nearest the camera. That is almost almost half of all the new cannon shots in the ditch. By contrast, nearly half of the new shot on the road are furthest from the camera.

More peculiarly, when considered by distance from the camera, nearly all new cannon shot on the road are further from the camera than any of the new cannon shot in the ditch.

Sixth, I note those cannon rounds that appear in OFF but that are 'missing' from ON. Ignoring 'missing' cannon shot from the edges of the OFF frame, at least three additional cannon shots in the center area of OFF are missing from ON. Granted, these are hard to spot without 'blinking' the two photos. One cannon shot is missing from the ditch, near to the center of the photo, at about a distance equal to the furthest 'new' cannon shot on the road. A second cannon shot is missing from the trail on the opposite side of the road, at nearly the same distance from the camera (that is, equal to the furthest new cannon shot on the road). A third is missing from the 'right side ditch' slightly closer to the camera. As can be seen by trying to locate these 'missing' cannon shot, each of these missing cannon shot are easily missed (or overlooked). While the new cannon shot in ON happened to come to a rest at 'strong' compositional locations, the only old cannon shots that were (re)moved, were (re)moved from very weak compositional points.

Finally, I would note a passage from Fenton's earlier letter describing the Valley of the Shadow of Death. I would assume that he is describing the location where he desired shooting the photos, but was prevented by ongoing Russian cannon fire. "Further on the balls lay thicker, but in coming to a ravine called the valley of death the sight passed all imagination [—] round shot & shell lay in a stream at the bottom of the hollow all the way down you could not walk without treading upon them . . .." Fenton's description places the shot and shell in the stream at the bottom of the hollow. The shot and shell collected in the lowest point, the stream, not on the road.

If I were to 'imagine' a scenario of what occurred 150 years ago, Fenton shot OFF, but was disappointed in not being able to shoot his stream with cannon shot and shell so thick you couldn't walk without treading on them. The ditch in OFF and ON, while it had collected a fair number of shot and shell, was a poor substitute. So, Fenton, after shooting OFF, decided to add some cannon shot to the foreground of the ditch. In short, he wanted to 're-create' the impact of his desired location. And then he got excessive and began placing shot on the road, where it would have an even stronger visual impact.

But that's just my two cents worth, and I could be as wrong as the next guy.

Let's apply our trusty Occam's Razor.

My rusty schoolboy history, and a little refresher from Google, recalls that the Russian artillery at the time of the Crimean War favored among other things, 12 pound cannon as an anti-personnel weapon. A 12 pound cannon fired a ball of, unsurprisingly, about 12 pounds weight of solid cast iron round shot. My guess is that's what we're looking at in these pictures.

It was a horrifyingly effective weapon of its day. A 12 pound, heavy, hard metal ball blasted at high speed from an artillery piece doesn't slow down until it hits the ground. The weight and speed equate to huge momentum. It passes straight through groups of people and horses unfortunate enough to be in it's path.

12 pounds (about 5.4 Kg) isn't very difficult to lift and carry, however. A brief examination of both photographs suggests there's about a couple of dozen more shot just on the road in the ON image. Looking at the surrounding ditches and paths, maybe that many more again?

If the ON picture is posed, we're being asked to believe that Fenton, perhaps aided by his assistant, scoured the out of view immediate hillsides collecting perhaps forty to fifty twelve pound round shot and deposited them throughout the nearer field of view. Not a physically easy undertaking. Possible by a couple of fit men, and perhaps if we discount the added hazard that they may have been under fire at the time. But a serious workout. Can this really be a reasonable explanation?

The alternatives are i) ON came first and shot was scavenged between the pictures, or ii) they were under fire, or close to it.

The Russian's British and French opponents also had plenty of 12 pound artillery, so the practice of salvaging round shot from the field was well employed on both sides. New shot was relatively expensive compared to that which could be picked up off the ground and especially compared to the labor of the lower ranks. Not all shot was salvaged. Misshapen or cracked shot would be left where it lay. Badly damaged salvaged shot could damage a cannon's barrel, or jam and cause a cannon to explode.

If I had to bet however, the most straightforward explanation is that ON was taken second. Fenton came under fire having first taken OFF. He realized the arrival of more shot made an even more striking picture, so took another.

In either case, I'd suggest that armchair psychological profiling of the photographer is unnecessary. Occam says that running around a hillside exhausting oneself collecting cannonballs, perhaps all the while risking getting cut in two by another one arriving, isn't particularly reasonable compared to the more likely explanations.

The comments to this entry are closed.