Sympathy for the Devil, part 1

This is a DRAFT. I’m publishing it in case some corrections might appear, but this has not been put into final form, necessarily.

In this case, the “devil” is the late Douglas Morrison. I have seen a cold fusion advocate express the opinion that Morrison is roasting in hell. However, of course, if Morrison is the Devil, he would simply be sitting beside a fire at home…. or in it, and not suffering, since Satan is made of fire.

I don’t wish hell for any human, because … what if I make mistakes myself? What if am as stupid as a flounder? We all can be seriously stupid when we become personally involved, the attachments make us stupid.

I’ve read that discussion before, but I found it confusing. Why? Idiotic ideas can be confusing, to be sure, but did Fleischmann’s arguments stand out as the soul of clarity, clearly visible against the backdrop of Morisson’s alleged  stupidity? Apparently not. Why not? Is it because Fleischmann was wrong? I don’t think so. I think there were other causes, and that’s what I will be looking for, here.

Jed Rothwell wrote, on

By the way, here is the most famous example of shooting a skeptical fish in a barrel. This is the debate between Morrison (who was roughly as stupid as a flounder) and Fleischmann. It would be hilarious if it were not so consequential. These are the best arguments the skeptics were ever able to muster. It is hard to believe but this was eventually peer reviewed and published verbatim. This demonstrates that idiotic ideas will fly through peer-review as long as they are opposed to cold fusion. There have been no other peer-reviewed critiques of cold fusion except Shanahan’s, which are even worse.

“Idiotic ideas will fly through peer review if …” If what? If they are opposed to cold fusion? No, that’s not accurate, at least not any more. Shanahan has had great difficulty publishing his critiques of cold fusion, he cannot make it through peer review. Rather, some ideas will make it through peer review if they confirm the opinions of the reviewers. Some reviewers might set aside their preconceptions — they are golden! — but many will not. It’s normal human behavior. To get past such a reviewer, one must establish rapport, connection, communication, one must “speak to the listening,” as it is called in my training. It is not necessarily easy, but if you think the reviewer is Satan incarnate, it’s impossible …. unless …. unless you can develop sympathy for the Devil. Which may or may not be advisable. The core concept of Satan is contempt and hatred for the human, and Satan wins when we have contempt and hatred for each other; in the ancient story, that is his goal, to demonstrate that humans are worthless junk, dirty garbage.

So, to that discussion.

The participants are Douglas Morrison and Martin Fleischmann. Fleischmann is the more notable of the two. Morrison’s opposition to cold fusion was important enough for his life that it was mentioned in his obituary in the CERN journal, specifically about his testimony before the Italian court in the libel suit filed by Pons and Fleischmann, a very nasty piece of business.

Morrison was a physicist, and the Pons Fleischmann experiment demonstrated an effect for which physics had no explanation. The leap to “they must be making some mistake” (I call that the Garwin Theory), then, was an easy one. Let’s look at the debate:

These two documents that first appeared in the Internet’s sci.physics.fusion forum in
1993. The first was written by Douglas Morrison (CERN), the second by Martin Fleischmann
(Univ. Southhampton) and Stanley Pons (IMRA Europe). Morrison wrote a critique of the
M. Fleischmann, S. Pons, “Calorimetry of the Pd-D2O system: from simplicity via
complications to simplicity,” Physics Letters A, 176 (1993) 118-129
Pons and Fleischmann respond to his critique.

According to a further note there, a similar paper was published in the ICCF-3 conference proceedings. That, 1992, would be this.

As published, hosted on New Energy Times. 

Morrison’s Physics A critique (paywall).

The Fleischmann and Pons reply (paywall)

Morrison’s 1999 assessment in Scientific American. 

This is the Morrison comment on sci.physics.fusion.


Newsgroups: sci.physics.fusion
Subject: Comments on Fleischmann and Pons paper.
Date: Wed, 12 May 1993 17:33:37 GMT
Organization: Sci.physics.fusion/Mail Gateway
5th DRAFT – Scientific Comments Welcomed. 6 May 1993.
Douglas R.O. Morrison.
M. Fleischmann and S. Pons [1] have published in Physics Letters A a communication entitled
“Calorimetry of the Pd-D2O system: from simplicity via complications to simplicity”. There they
claim evidence for the production of excess enthalpy of greater than one kW per cc of Palladium
in a Pd-D2O system. They comment that this is comparable with the rates obtained in a fast
breeder reactor. They note that the reproducibility is high. In this letter serious doubts are
expressed about this claim and the methods used to derive it.

Indeed. First of all, the “fast breeder reactor” comment. Did F&P  actually write that? From the abstract:

We present here one aspect of our recent research on the calorimetry of the Pd-D20 system which has been concerned with high rates of specific excess enthalpy generation ( > 1 kW cm^-3) at temperatures close to (or at) the boiling point of the electrolyte solution.

And then, p. 128:

We note that excess rate of energy production is about four times that of the enthalpy input even for this highly inefficient system; the specific excess rates are broadly speaking in line with those achieved in fast breeder reactors. We also draw attention to some further important features: provided satisfactory electrode materials are used, the reproducibility of the experiments is high; following the boiling to dryness and the open-circuiting of the cells, the cells nevertheless remain at high temperature for prolonged periods of time, Fig 8; furthermore the Kel-F supports of the electrodes at the base of the cells melt so that the local temperature must exceed 300ºC.

There is no “nuclear” claim in the F&P paper. However, by this time, a physics audience was already sensitized, and the claim would be hanging in reader’s space. It was completely unnecessary to mention breeder reactors. Even though it was true, as to their measurements, it would increase immediate skeptical reaction. With that reaction, the evidence presented will be mentally filed in a different way than with pure heat claims. Now, was this “reproducible.” Notice the qualification, a loophole one could drive a truck-full of palladium through: “”provided satisfactory electrode materials are used.” The fact that could be reported would be more like “with such and such materials, we found this effect to be reliable in our lab. Let someone else claim reproduction, not “reproducibility,” which is always a bit of a fantasy, until and unless work has actually be replicated. So no wonder Morrison doubted. It was a setup, and it likely caused him to disregard the substance of the paper.

Essentially they perform electrolysis in a transparent test tube which is open so that the gases
and vapour can escape freely. The cathode is a small rod of palladium of 0.2 cm diameter and
1.25 cm length giving a total volume of 0.039 cm3. There are three stages.

A “transparent test tube”? No. That misses that the cell is a half-silvered “single compartment vacuum Dewar.” Morrison is a physicist without extensive experience with calorimetry and easily can miss these details. As well, the gases and vapor, whatever vapor there is, cannot “escape freely,” as I would imagine that. With that, I imagine an open top. Instead, there is a very narrow neck, a long tube, through which gases may escape (or, theoretically, enter, though that would largely not happen, and this becomes important later, I think.

For the first stage a moderate current is used for electrolysis. It is noted that at short times the
heat transfer coefficient decreases – this they ascribe to the heat of absorbtion of hydrogen ions in
the lattice.

This probably needs more attention. Initial loading is exothermic. As loading increases, it begins to require energy to load the deuterium into the palladium. Being able to detect these effects is, if I’m correct, a sign of the precision of FP calorimetry, which is reputed to have been the best in the world, at the time, with a precision of about 100 microwatts, at least in some of their work.

In the second stage the current densities are raised to increase the temperature above 50 C – this
with D2O. Finally, in stage three, the cells are driven to boiling point. A complicated (non-linear
regression) analysis [2] is employed and it is calculated that there is excess enthalpy generated in
the lattice, the amount calculated increasing steeply with time (and temperature).

I would not have raised the temperature this way, it’s messy, and I always reacted to the boil-off experiments with some level of skepticism. Rather, I’d have suggested using supplemental heat, which can, of course, be measured precisely, to take the temperature to boiling. Recent experiments by Storms have suggested that the electrolysis current is not necessary, once the reaction has begun, to maintain it. By using electrolysis power to heat the cell, this effect was concealed, which is unfortunate.

Morrison complains that the analysis is complicated. This would be a common complaint. How complicated is it, for a paper published in PLA?

However, this is my reaction to the analysis. My eyes glaze over. It is meaningless to me. To understand it would take work. Now, imagine I’m a physicist reading this physics journal. I am not familiar with the FP calorimetric analysis. Will I do the work to understand it, if I already suspect that this is complete BS?

For me, I am intensely aware, this is hindsight analysis. Nevertheless, this is what I’d have suggested: the calorimetric method would be published as a method, and it would be thoroughly calibrated. With calibrations, the complication of the analysis is actually bypassed. Calibration proves the method, demonstrating its accuracy and precision. Then, in a paper reporting anomalous heat, all that would be in the paper would be heat results, plus some calibrations that show the precision and accuracy under these experimental conditions. It is often just a single plot. It can be quickly and easily uinderstood. But … this was not done, and FP papers were written for those willing to put in major effort.

In the third stage the behaviour near and during boiling is observed using a video camera.
From this video, the time for the cell to go from about half-empty to dry, is taken – more
precisely the amount of liquid boiled off is estimated over the final 10 minutes before the test
tube was declared dry. A new simple calculation is made in which the enthalpy input is
calculated as
(cell voltage – 1.54 V).(cell current)
and the enthalpy output is taken as composed of two terms, the energy radiated and the heat
resulting from the vapourization of the D2O remaining in the cell 600 seconds before it is dry
(this latter term is dominant). It is this simple calculation that gives the highest values claimed,
namely “the excess rate of energy production is about four times that of the enthalpy input” and
that the excess specific rate is 3.7 kW per cc of Palladium.

As I read this, the D2O in the cell evaporated much faster than expected from heat input. Now I understand, for the first, time, the title of the FP article. The initial calorimetry was complex. This was simple, relatively speaking, or so it could seem. So along comes a physicist to teach these chemists some chemistry.

There are several major problems with this calculation.
First is that the “cigarette lighter effect” has been forgotten.

My immediate reaction: the arrogant asshole! Did these chemists forget about release and recombination? That would be amazing! It would be something they would be intimately familiar with.

In the last century it was difficult to make reliable matches to light cigarettes. A reliable smokeless lighter was invented which consisted of a rod of palladium into which hydrogen had been introduced under pressure. This caused the lattice of the palladium to expand and thus stored energy. To light a cigarette, the top of the rod was uncovered; some hydrogen escaped releasing some of the stress and thus releasing energy which resulted in a small rise in temperature of the end of the rod. Palladium is a catalyst of hydrogen and oxygen which burn to give water plus energy. The palladium now slightly heated, catalyzes the escaping hydrogen and the oxygen of the air and the resulting heat of combustion which is mainly deposited on the surface of the rod, raises its temperature. This temperature rise releases more hydrogen which is catalyzed by the still more efficient hot palladium, and so on until the tip of the rod is so hot that the cigarette can be lit. The reliability of this system is high.

Realize that Morrison is writing for a naive audience. He is creating an artifact which is plausible to his audience, which is not familiar with the experimental conditions. Yes, palladium catalyzes the recombination of hydrogen (or deuterium) and oxygen. In Miles’ work with codeposition, small pieces of the palladium plating on his cathode broke off and would deload, floating on the surface of his electrolyte, and they were flaming, because there was oxygen in his headspace, as is normal in these electrolytic cells, when they are operating with electrolytic current. However, the palladium cathode will store a large volume of deuterium gas. Meanwhile, the evolved oxygen escapes from the cell, only the recently released oxygen would be there, a limited quantity. So when the palladium rod is exposed to the cell gas, is no longer immersed in the electrolyte, it can then catalyze recombination. There will indeed be some heating. However, this heating will be limited by the available oxygen, and this is where that narrow opening becomes more important. If gas is being released from the palladium, which will start to happen immediately with highly-loaded palladium (as this would be), it and any generated water vapor (the product of such combustion) will create a flow out of the cell, maintaining the flow that was already there. When the D20 is boiled off, no more oxygen will be available and recombination heat will cease, but deuterium will continue to be released, for quite a long time, actually. At lower loading, this is an endothermic reaction, it is only at high loading (under pressure, as Morrison puts it) that release is exothermic.

As well, the amount of heat available this way is low, compared to what was measured.

In the simple calculation used for stage three, a significant effect is omitted, of the heat
produced by the catalytized recombination of the hydrogen with the oxygen. The oxygen is
released from the anode by electrolysis, and towards the end when the cell is about dry, from the
air. There is no mention in Fleischmann and Pons’s paper of any attempt to measure the amount
of oxygen, deuterium and water in the gases and vapours leaving the test tube.

Bottom line, this indicates to me that they did not expect this critique, because it would all be obvious to them. There would be some recombination, transiently. How large an effect would it be? Morrison does not look at this quantitatively. Morrison continues to assume plentiful oxygen.

In the Fleischmann and Pons paper, it is noted as a further demonstration result, that
“following the boiling to dryness and the open-circuiting of the cells, the cells nevertheless
remain at high temperature for prolonged periods of time (fig.11); furthermore the Kel-F
supports of the electrodes at the base of the cells melt so that the local temperature must exceed
300 C.” This dramatic effect cannot be explained by Fleischmann and Pons as being due to
electrolysis since there is no liquid and no electrolysis. However it is exactly what would be
expected with the “cigarette lighter effect” where the hot palladium rod continues to catalyze the
interaction of the hydrogen which is slowly escaping from the rod, with oxygen from the air.
It might be expected that this effect would occur also with normal water, H2O, being used
instead of heavy water, D2O, but no description is given in the paper of any results of tests of the
stage three boiling using normal water, H2O.

This is a variation on “they did not do the control experiments we want to see.” Or, perhaps more accurately, they did a lot of work that has not been reported. There is no doubt that such a control experiment would have been of interest. I’ll be interested to see how F and P respond to this. Given the situation with oxygen, though, the melting of the cell supports would not be expected. The actual analysis could be somewhat complicated, because of the messiness of the boil-off process, my analysis. But it could be done, and, of course, measurements of the cell gases would be of high interest as well.

An interesting confirmation of this using electrochemistry was reported by Kreysa, Marx and
Plieth [3]. They write “We have to report here that as we removed the deuterium-loaded
palladium sheet from the cell and laid it on the table it did burn a scald into the table. One can
still argue that this was due to deuterium fusion.

Oh! Nonsense! Really, this is embarrassing, I am starting to agree with Jed. Sure, one could argue that, stupidly, but this was deuterium-loaded palladium exposed to air, with plenty of oxygen. The palladium will release deuterium which will, at the surface, be recombined with oxygen, and while the release itself is endothermic, the recombination overcomes that, it is absolutely no surprise that the sheet would become hot. But inside an enclosure, where the released deuterium — slowly, perhaps over days — keeps air from entering, no, it would be very surprising. Generally, the sheet would cool.

Morrison then gives numbers, details, for this heat, as if it were relevant. I’m not quoting that. What would be needed would be the experimental behavior of a palladium cathode under conditions where no XP was being measured (a “dead cell” experiment), when boil-off occurs. What did F&P do?

[…] It may be noted that Fleischmann and Pons used an exceedingly small piece of palladium,
0.04 cm3, which works well as a catalyst, but which means that after catalyzing a larger volume
of heavy water, the power calculated is apparently larger than with Kreysa et al. because the
volume of palladium is so small.

What? The volume of heavy water “catalyzed” is irrelevant. What is relevant would be the amount of deuterium loaded into the palladium, and then the release rate. The palladium will only hold so much deuterium. Before Pons and Fleischmann’s work, it was thought that loading above 70% (7 deuterium atoms for every 10 palladium) was not possible. It was, and it is possible that some local loading above 100% may be possible, but above 100% is rare as a claim.

Should Drs. Fleischmann and Pons wish to test their previous
conclusions [1], it would be interesting if they were to describe experiments where they repeated
their published experiment but with a substantially larger amount of palladium and a relatively
small volume of D2O.

Weird. It is very difficult to set up FP Heat Effect conditions. You can’t do just anything. I’m beginning to see stupid  as a flounder having legs. However, that observation is not particularly useful. Morrison had legs, flounders don’t. The useful question would be how to respond to critique like this in the future, and that is pretty much the only serious utility to this examination. “So and so was stupid” is not an empowering argument. Rather, understanding what led Morrison to what he shows here would be more useful. Obviously, these critiques convinced him, and they also passed peer review, though peer review of Letters may be poor, we have seen that. Jed claims this was the best the skeptics could come up with.

Pons and Fleischmann deliberately scaled down, because of that early meltdown, which, combined with the erratic nature of the effect, it being difficult to predict whether a particular experiment would generate heat, but even more, how much heat, made scale-down a necessity. Attempting to convince skeptics by scaling up would be a Bad Idea. It wouldn’t work, anyway, very likely.

Secondly, there is the assumption that ALL the liquid present in the tube 600 seconds before
dryness, was boiled off. That is none of it was carried out as a liquid, from the test tube. Now the
video shows that there is highly turbulent motion. And as Kreysa et al. [3] showed, 74 seconds
after the palladium becomes dry, temperatures of a few hundred degrees can be reached. Thus it
is reasonable to expect that with such a chaotic system, some fraction of the liquid is blown out
of the test tube as liquid and therefore should not be counted. The existence of such a fraction is
omitted from the simple Fleischmann and Pons calculation. And no attempt to measure this
fraction is described.

This is a valid objection, though, i.e., that volume should be measured. It could also be estimated through various techniques. If electrolyte is blown out, it is not just D2O, it contains lithium sulfate. Even if it’s hot and vaporizes immediately, the lithium sulfate would remain and would deposit on surfaces. It has been pointed out by many researchers that one cannot state everything in a paper. The issue would be what’s important. To these electrochemists, they knew there wasn’t substantial recombination, and they knew that their electrolyte wasn’t being blown out that exit tube. I’ve seen this skeptical argument many times, but it would generate visible effects, especially deposits of cell salts outside the exit tube. We will see that F&P mention the finding of lithium outside the cell. It’s an issue of quantity.

Thirdly, the input enthalpy is taken as the current multiplied by the (cell voltage – 1.54V). It is not explained how these quantities are measured. This is crucial as when the cell is boiling vigorously, the impedance must be fluctuating strongly. Thus the current will have both an AC and a DC component. If only the DC component were measured, then the input enthalpy would be underestimated. A detailed description of the current and voltage measuring systems showing their fast response characters is needed, but is not presented.

I’ve seen another pseudoskeptic come up with this argument. Again, not everything was stated, for sure. I don’t know what systems F&P were using, but SRI used a constant current DC supply which regulates current with a bandwidth of at least 100 kHz. So they set current and recorded voltage, with voltage samples being averaged. Bubble noise does exist, but is slow by comparison. Dieter Britz did an analysis of bubble noise and the effect on power measurements, and the effect was very small, not significant. Further, the power input goes to zero when the cell boils off, but heat generation apparently continues. Morrison ascribes that to recombination, but there isn’t enough oxygen there for that.

Since these three important aspects of the experiment have been omitted, it is not possible to
say whether or not excess enthalpy has been observed in the last 600 seconds to dryness (stage

I will be coming to this point: Morrison confuses it being impossible for him to say whether or not what was observed was excess enthalpy. He obviously either doesn’t have enough information, or has misunderstood what information he has. And this is why we look for independent confirmations and, to declare artifact, rather than the mere possibility of some artifact, we look to controlled experiment that identifies the artifact, i.e., that reproduces the reported effect and then demonstrates the origin. However, for many, it was enough that some doubt could be raised. It was missed that the 1989 DoE review recognized the difficulties and did not declare that cold fusion was dead, merely that conclusive and confirmed evidence was not known  to them.

I would never say that some critic should accept the reality of cold fusion, unless personally convinced. However, with some critics, it becomes obvious that the critique is driven by something other than clarity and an objective seeking of reality. Morrison does not here claim that Pons and Fleischmann were wrong, only that possible doubts of his have not been resolved.

There are two important problems with stage two. Firstly, a complicated non-linear regression analysis is employed to allow a claim of excess enthalpy to be made. This method of Fleischmann and Pons [2] has been carefully studied by Wilson et al. [4] who state that “they significantly over-estimate the excess heat…….an additional significant overestimate of excess energy occurs when the calibration is made above 60 C”. Now stage two is mainly above 50 C and rising to 100 C. Further Wilson et al. write “Because of the paucity of experimental details in their publications, it has been difficult to determine quantitatively, the effect of calibration errors.” A reply by Pons and Fleischmann [5] did not address the main questions posed by Wilson et al.

I would need to study Wilson and the P&F reply to comment on this. However, this is merely stating that there were unaddressed questions.

Secondly, it may be noted in fig. 8 of ref 1, that the cell voltage rises as the temperature rises and that as 100 C is approached, the voltage rises more and more steeply. Experience by the GE group [6] was that in operating similar open cells over many hours, they also noticed a rise in cell voltage with time. They attributed this effect as being due to some of the escaping gases carrying some Lithium with them. As the level of the electrolyte is maintained by adding fresh D2O (but not any lithium salt), the concentration of lithium in the electrolyte decreases with time and the voltage rises. This was proved by atomic absorption analysis, that the cell resistance had risen (causing higher voltage due to the constant current mode operation) due to loss of lithium which was caused by sputtering of electrolyte droplets up the gas outlet tube. This may be considered confirmation that even at moderate temperatures, the outlet stream contains liquids as well as gases as discussed for stage three when the temperature was much higher and the boiling much more vigorous. It may be concluded that claims of excess enthalpy in stage two have not been established.

This is similar. Any electrochemist would be thoroughly aware of these possible issues. There are other possible causes for an increase in resistance, such as anode passivation. There are rate issues here. Bottom line, not everything was covered in the F&P paper. The GE group was using, I assume, a different cell design, it may show differing phenomena.

The overall conclusion is that many important factors have been neglected so that it has not been established that excess enthalpy was observed.

Are the “neglected factors” — allegedly neglected, more accurately not discussed — important? That is a question that would require quantitative analysis, not merely a qualitative assertion of possible relevance, and it would require much more communication. We will see how Pons and Fleischmann respond, in the next post

The experiment and some of the calculations have been described as “simple”. This is
incorrect – the process involving chaotic motion, is complex and many calibrations and
corrections are needed.

Morrison is quibbling about semantics here. The boil-off period is chaotic, yes, but is conceptually simple, because input power ends when the electrolyte is boiled. To neglect that heat, Morisson depends on the “cigarette lighter effect,” which is implausible under the experimental conditions. Could this be better explored and explained? I’m sure. But “simple” is not wrong.

The calculations have been made to appear simple by incorrectly ignoring important factors. It would have been better to describe the experiments as “poor” rather than “simple”. A true “simple” experiment is one where corrections and calibrations can be reduced to a minimum. This can be achieved in calorimetry by using a closed cell and by enclosing the cell in a series (eg three) baths which are each kept at constant temperature. The
cell is kept at a higher temperature than the innermost bath so that if any excess enthalpy is produced, the heating of this bath can be reduced thus measuring simply the excess. Since this is a null measurement system, there is little need for complicated corrections. It is to be regretted that in the nine and a half years (the last four years well-funded) that Fleischmann and Pons say they have been working on this [7], that they have employed such a simplistic open-cell system.

If those factors that were not covered are not major factors, then the calculations were indeed simple, and correctly so.

It is a pleasure to acknowledge the help of many friends, in particular D. Britz, F. Close, T. Droege, R. Garwin, and S.E. Jones.
[1]. M.Fleischmann and S. Pons, Phys. Lett. A 176 (1993)1.
[2]. M. Fleischmann and S. Pons, M.W. Anderson, L.J. Li, and M. Hawkins,
J. Electroanal. Chem. 287(1990)293.
[3]. G. Kreysa, G. Marx, and W. Plieth, J. Electroanal. Chem. 268(1989)659.
[4]. R.H. Wilson, J.W. Bray, P.G. Kosky, H.B. Vakil, and F.G. Will,
J. Electroanal. Chem. 332(1992)1.
[5]. M. Fleischmann and S. Pons, J. Electroanal. Chem. 332(1992)33.
[6]. General Electric group of ref. 4. priv. comm.
[7]. Press release, University of Utah, 23 March 1989.

I see something basic here. The FP paper appeared in a physics journal, not a chemistry or electrochemistry journal. Morrison was not qualified to review this work. The review should have been done by experts in the relevant field, otherwise many errors could be made. There is no “physics” claim in the FP paper. It’s all chemistry and heat, and the critique is essentially chemistry. Why did F&P publish in APL? That, to my mind, more or less invited a response like this.

Next part: Fleischmann and Pons reply.

Author: Abd ulRahman Lomax


One thought on “Sympathy for the Devil, part 1”

  1. Dear Abd,
    I met Morrison at ICCF-2 Como, he told that initially he believed that CF woks but was disappointed by the results- their quality. Not a sadist or an enemy by definition, anyway.
    See please
    – my paper FROM 317…(joke)
    He remembered me because due to isolation and oppression in 1990 I did not know what an
    e-mail is- for real/unreal.

    An other personality I met at Como was Heinz Gerischer a great electrochemist who was very nice but confessed he does not know what to think about Cold Fusion. Had no certainty.


Leave a Reply