I find Le Conte's reasoning interesting. It's almost seductive even now. He's absolutely right that we don't have systems which approach the efficiency of the "animal machine".
But that efficiency is due to our remarkable performances on nothing but a few grams of low-grade fuel. It doesn't say anything about what would happen if you had some really high-grade fuel.
Sugar has about 18.8 joules per gram. Kerosene, which was pretty well known in 1888, has almost 46 000 joules per gram.
Perhaps we can extract some general principles about prediction from that, like: given a complex enough system, if one factor changes by two or more orders of magnitude, previously observed behavior is useless as a guide to the future. I'm fudging on 'complex enough', though. Trigonometry works just fine with triangles the size of pencils or skyscrapers, but predicting the scope of inventions doesn't seem to work well when one factor changes radically.
The fudge point in his reasoning: "The animal machine is far more effective than any we can hope to make."
Exactly what does effective mean?
We still can't make ornithopters that approach the performance of birds. That is what his analysis actually shows. Our successful aircraft still don't have performance that resembles that of birds. Then again, no birds cruise at an altitude of 30,000 feet at nearly Mach 1. Our aircraft don't have to be bird-like to be economically useful, and this is indeed effective to us.
So to go back to my other comment that predictions should involve economic analysis: our professor should have considered flight performance envelopes that would be commercially useful, not flight performance envelopes that resemble those of birds. (For example, he could have extrapolated something like the performance envelope of a balloon, but with twice the speed and better directional control.) If he had considered a few examples of those, he'd have quickly realized that using birds as a model wasn't a good idea for an analysis of human flight technology.
The 18.8 number is kilojoules / gram (food energy).
The real limit is our ability to burn huge amounts of it. Even a small plane burns over 1 lb / minute, which would be a shocking metabolism for an animal.
Damn, sorry about that, and thanks for the correction. So the real difference between petroleum derivatives and sugars isn't the energy content so much as the power you can generate? In layman's terms, the kablooie factor?
That puts Le Conte's error in even more perspective. Imagine if you said to him in 1888 "...but what if you had access to vast, cheap supplies of high-grade fuel which you burned off at a rate of 1 lb per minute?" He'd look at you like you were an idiot. Now your heavier-than-air craft is carrying several hundred pounds of fuel too, just to stay aloft for a few hours? And what do you do when you get where you are going?
Does this mean Le Conte's real failure isn't energy calculations, but predicting the availability of such fuel? In other words, it's all economics?
I had another comment about "toys" versus "tools for real work" in this thread. Maybe that's another defining element of "toys", the difficulty of maintenance and the lack of ubiquitous infrastructure. Only rich or eccentric people are going to take the trouble to provide infrastructure for their own machines, so it becomes by definition a "toy". Like planes or cars in the early 20th century.
Okay, is this a new heuristic? We look for devices which are fun or useful but impractical because right now, enthusiasts have to provide their own infrastructure at great cost?
But that efficiency is due to our remarkable performances on nothing but a few grams of low-grade fuel. It doesn't say anything about what would happen if you had some really high-grade fuel.
Sugar has about 18.8 joules per gram. Kerosene, which was pretty well known in 1888, has almost 46 000 joules per gram.
Perhaps we can extract some general principles about prediction from that, like: given a complex enough system, if one factor changes by two or more orders of magnitude, previously observed behavior is useless as a guide to the future. I'm fudging on 'complex enough', though. Trigonometry works just fine with triangles the size of pencils or skyscrapers, but predicting the scope of inventions doesn't seem to work well when one factor changes radically.