Administering exogenous insulin as a therapy for insulin dependent diabetes ( type 1 ) is a mixed success. The problem stems from the fact that while the therapy address the endocrinology correctly, there are currently intractable anatomical problems with it. If we were to wager as to where any randomly selected insulin injection was administered, the smart money would be on the abdominal fat pad as the most frequent injection site. Lesser used, but still observed, injection sites are the arms, legs, and buttocks.
All Roads Lead To Rome
These injection sites pose a problem which is immediately apparent if you take a look at a diagram of gastrointestinal tract blood flow:
The thing to take away from this diagram is that when the body self-administers insulin via the pancreas, it does so directly into the pancreatic veins, which after a very short run flow into the lienal vein, which, in turn, flows directly into the portal vein. So, the primary client for pancreatic insulin is ... the liver. In fact, the majority of secreted insulin does not initially make it into general (peripheral) circulation as it is bound to hepatic receptors. While this fact is interesting in and of itself, the slightly more subtle point is that the pancreas and liver are a superbly coordinated tag team when it comes to maintaining normoglycemia.
This observation helps us to explain your blood glucose patterns. Prior to your first meal of the day, your body is in a post-absorptive fasted state, meaning that your liver is busily ensuring that blood glucose does not dip too low by engaging primarily in glycogenolysis, gluconeogenesis, and ketogenesis. When your first potato meal appears in the portal vein, your liver is forced to transition from glucose release to glucose absorption. This is part of the liver's job description, so it will do so willingly, without complaint. It will not, however, do this instantaneously, taking on the order of about 30 minutes or so to shift metabolic gears towards absorbing glucose and glycogenesis.
It is due to this delayed switch to glucose uptake from the liver that more of the ingested carbohydrates will initially be able to make it into peripheral circulation where you can measure them with your glucose meter. By the time you have your second potato meal, the liver has already amortized the cost of transitioning to glucose absorption, so it reacts much more quickly to the incoming glucose and insulin in the portal vein, which is reflected in the lower blood glucose numbers that you observe.
I find your comments relating to the ketone measurements interesting and somewhat perplexing since ketone production is an indication of lipid oxidation in the liver, and the liver generally does this under starvation conditions, or conditions mimicking starvation, i.e. when there is a shortage of carbohydrates, which is certainly not the case when you are only eating potatoes.
Now, since you did ask me for my thoughts, let me give you some raw stream of consciousness speculation. My sense is that if what you observe regarding oxidation of fats is typical, it may be the case that what you are seeing is the result of inducing an essential fatty acid deficiency.
As far back as 1929 in a study entitled "A New Deficiency Disease Produced By The Rigid Exclusion of Fat From The Diet", it was known that there are classes of fatty acids that are essential for the normal development of rats. The poor rats fed this fat free diet developed a number of issues, but the thing of interest to us is that at the time of death, none of the rats had appreciable body fat depots. In fact, in the summary table, upon autopsy, many rats were reported to have "no body fat" at all:
In this study, rats were weaned on an initial diet of 24% protein / 76% carbohydrates, and then in the maintenance phase, protein was brought down to 12%. It just so happens that a baked potato is about 90% carbohydrate and 10% protein by dry weight. Of course, these rats spent significant portions of their lives eating this way, so we can't consider this to be directly comparable to your eating potatoes exclusively for a couple of weeks, but while this is not probative, it does lend some support to the notion that a diet devoid of dietary fat can foster significant adipose tissue oxidation, albeit under pathological conditions.
Another study, however, that involved human participants being administered diets of 80% glucose and 20% amino acid hydrolysates is much more relevant to you. This study only lasted two weeks, and this is what the authors had to say:
So, it only takes about 3 days to fully manifest essential fatty acid deficiencies on a diet completely devoid of fats. One major caveat here is that these subjects were on a constant glucose infusion which suppressed endogenous FFA release entirely. This is important, because researchers calculated that subjects had sufficient adipose stores of EFAs to last in excess of 8 months, if only they could be liberated.Originally Posted by Wene et al.
With all of this as background, my speculation is that during the potato diet, your body turns to aggressively oxidizing fats not due to an energetic imperative, but rather because it needs to access the stored EFAs in order to synthesize the biochemically critical eicosanoids and prostanoids, compounds that are pivotal in mediating numerous physiological and biochemical processes.