This website contains other cold fusion items.
Click to see the list of links

373) The Curie Project

Ludwik Kowalski

Montclair State University, New Jersey, USA
June 22 , 2009

1) Introduction
History of my CR-39 cooperation with Richard Oriani is summarized in unit 369. The beginning of The Curie Project is described in unit 368. At the end of this unit I plan to show results from that cooperative project; I hope they will be available in September, before ICCF15 (15th International Conference on Cold Fusion in Rome, Italy).

As indicated in the title, this unit was created for reporting anticipated experimental results. But I changed my mind; results will be shown in a later unit. What follows are selected messages posted on the private Internet list for CMNS researchers.

= = = = = = = = = = = = = = = = = = = = = = =

2) Another appeal to CMNS researchers
Here is a message I posted (6/24/2009):

‘1) As I wrote earlier, The Curie Project is in progress. But it has only three participants (from the Phys-L list for physics teachers), plus myself. We are using Oriani's protocol; it is described in his ICCF14 contribution.

The only two departures from the protocol are: 

a) Keeping chips in salty water, when they are waiting for electrolysis and when they are waiting to be etched. Richard wrapped them in Al and kept them is air.

b) The rear side of a chip, during electrolysis, is protected from being in contact with air (by a ~0.5 mm layer of polyethylene). Richard did not protect the rear surface from air. The project desperately needs more participants; it is not too late to join us.

2) Here is a summary of Oriani's results; this does not include nine recent experiments. These experiments will be counted as his contribution to The Curie Project. Richard is using the same CR-39 I am using (we bought one Fukuvi sheet together, two years ago). Not surprisingly, his mean background, measured 13 times, is nearly the same as my background, measured 6 times (so far). His mean background was 13.6 tr/cm2; his standard deviation was 6.8 tr/cm2 (as shown in Table I).  My mean background was 13 and my standard deviation was 5. 

a) Richard’s post-electrolysis chips had tracks on both surfaces. His Table II shows results from 22 experiments (#4, #5, #6, . . . . #25). As you can see, results differ very significantly from one experiment to another. But 40 out of 44 surfaces had track densities higher than 16 tr/cm2. In other words the number of post-electrolysis tracks exceeded the number of background tracks in 90% of cases. That is remarkable.

b) Let me summarize Oriani’s results in terms of mean values and standard deviations.

Front surface (facing the electrolyte) mean=142; st. dev.=146
Rear surface (facing away from the electrolyte) mean=132;

The lowest density was 9 and the highest density was 498 tr/cm2. In one case the number of tracks was too high to count; I tentatively turned this into 498. The two distributions are very wide. But the mean track density on each side of the post-electrolysis chip, is about ten times higher than the mean density on control chips. This includes the four cases in which the densities were only 9, 9, 11 and 16. The distribution of track densities, on 44 surfaces of post-electrolysis CR-39 chips, is shown in Figure 1 below.

Figure 1 (373histogr.png)

Why is the histogram so strongly skewed? I am not at all comfortable with the fact that the overall mean, 147, and the same as the standard deviation. Yes, 147 exceeds the background by one order of magnitude. But the level of confidence that this would be confirmed in another set of 22 experiments is very low. I was trained to believe that, in order to be taken seriously, a difference between a signal and a background should be at least as large as the sum of standard deviations (of the signal and the noise), preferably two or three times larger. On the other hand the above curve shows that densities exceeding the background are not at all rare. Is there a better way to evaluate a level of confidence than standard deviations? ‘

3) Categorical data
Suppose Richard’s data are used to answer simple “yes” or “no” questions. The questions would be:

Is it true that track densities are higher than mean background?
Is it true that track densities are higher than 69 tracks/cm2?
Is it true that track densities are higher than 140 tracks/cm2?

Richard asked the first question 44 times and the answer was “yes, it is true” was obtained 40 times. On that basis one would be able to say that f=40/44=0.91 represents the expected probability that the statement is valid. In other words, 91% of CR-39 surfaces, in another sequence of identical 22 experiments, are expected to validate the statement (that the signal is higher than the background).

This kind of predictions is common. Here is a typical illustration. Suppose a random sample of n=10 people was selected in a town. Six of them speak Spanish. In this case the answer is either "yes" or "no." A fraction f=6/10=0.6 is calculated on the basis of this small sample. How reliable is the statement that "about 60% of people, in this town, speak Spanish? "  The answer depends on the value of f, and on the size of the sample. The distribution of f, from many samples of size n, is expected to be nearly Gaussian, when n>20 (actually, a it is already nearly Gaussian at n=10). The width of the distribution of fractions (actually the standard deviation of f) is expected to be equal to sqrt [ f*(1-f) / n]. 

The uncertainty about the expected 91% reproducibility, in subsequent sets of similar 22 experiments, the standard deviation s=sqrt [ f*(1-f) / n] = sqrt [0.91*0.09/44]=0.043. Subsequent  sets of experiments are expected to produce values of f that are between 0.91-0.043=0.877 and 0.91+0.043=0.953 (two standard deviations). The theoretical reproducibility of nearly 91%, based on real experimental results from one set of 22 experiments, is a rather convincing argument for the claim. That "more than background" claim, however, is not very strong. 

So let us make a stronger claim, for example "more than 69 tr/cm^2, as above." Richard reported 25 such outcomes out of 44 measurements (f=25/44=0.57). Thus only  57% of sets of replications are expected to support this new claim. The predicted reproducibility for evan stronger claim, "more than 140 tr/cm^2" is considerably lower than 57%. All this is intuitively obvious; stronger statements, based on the original set, are less likely to be supported (by results from subsequent sets of 44 measurements) than weaker statements. But thinking in terms of standard deviations of fractions might be helpful.  I am learning this from an elementary statistics textbook. According to the "central limit theorem," the distribution of mean values, from several sets of experiments, is Gaussian, even when distributions from individual sets are not bell-shaped.’

4) Ludwik’s message posted on CMNS list (6/28/2009)
“The attached histogram [see Figure 1 above] is based on Richard's Table II in

It has 11 bins and bins are populated with the following numbers of observations:

14, 10,, 6,  3,  3,  1,  1,  1,  1,  1,  1

The purpose of The Curie Project is to verify Richard's results. He examined 42 surfaces and measured track densities on them. Suppose several people also measure track densities on 42 surfaces, from their own sets of experiments. The histograms would be quite different. This is intuitively obvious, considering small numbers of observations (see above). I verified this by simulating replications with the Monte Carlo method. To accomplish this I simply assumed that the true distribution (one that would be obtained if thousands of surfaces were examined) is the same as Richard's distribution. Below are results from my first five simulations. Each line is a set of numbers to plot one histogram (and to calculate its mean value). The last numbers, after the asterisk, are mean track densities for the five histograms.

19,  7,  8,  8,  0,  0,  0,  0,  0,  0,  0  * 106
14, 16,  2,  3,  4,  3,  0,  0,  0,  0,  0  * 121   
9, 11,  7,  7,  4,  0,  0,  0,  3,  0,  1  * 164
21,  7,  0,  8,  3,  0,  0,  3,  0,  0,  0  * 126
11,  8,  6,  0,  2,  0,  4,  0,  0,  1,  3  * 187

The distribution of mean density, from 1000 simulated histograms turned out to be bell-shaped; the mean of means was 154.5 tr/cm2; the standard deviation was 18 tr/cm2. Individual histograms are strongly skewed but, according to the Central Limit Theorem, the distribution of the mean of means should be Gaussian. My Monte Carlo simulations confirmed this. And they provided a qualitative definition of "success,"  for The Curie Project.

The purpose of that project is NOT to replicate the shape of Richard's histogram; it IS to show that the mean track density, on post- electrolysis chips, is not very different from his value of ~140 tr/ cm2. What does it mean "not very different" ? The difference should not be larger than two standard deviations (2*18=36). In other words, any mean density, between 140-36=104 tr/cm2 and 140+36=176 tr/cm2, would be consistent with Richard's result, provided tracks on many surfaces are counted in The Curie Project. That is why additional participants are needed. It is not too late to become a participant. Comments and suggestions will be appreciated.”

5) How to mount electrodes

The attached figure shows (very schematically) how to make a cell without a titanium rod in a guiding glass tube (see Oriani's illustration). My cell is open and the rigid column slides into it.

Figure 2 373column.png

6) My CR-39 Chips for measuring background

Each of my post-electrolysis chips has a matching control chip--it is kept in the unused electrolyte during electrolysis. In additions to this, one control chip will be mounted inside the electrolytic cell in exactly the same way as it is during electrolysis, including the position of electrodes. But electrodes will not be connected to the power supply, as Richard did. This test will be much longer than three days; it will show me what is the total contribution of radioactivity, if any, from the polyethylene, Mylar, o-rings, the electrolyte and the cathode.

7) Another message posted on CMNS list (7/2/2009)
“As I specified before, there are only two departures from Richard's protocol,


in the ongoing project. They are:

(a) Keeping chips in distilled salty water, when they are waiting for electrolysis and when they are waiting to be etched. Richard wrapped  them in aluminum and kept them in air. During each electrolysis his control chip is kept in the unused electrolyte. We am doing the same thing; in the past (see my Catania report) I kept control chips in air.

(b) The rear side of a chip, during electrolysis, is protected from air (by a ~0.5 mm layer of polyethylene).

Why am I trying to minimize exposure to air?  Because I suspect that electrostatic charges somehow appear on CR-39 surfaces kept in air. Some areas become positive and others become negative. The reason for suspecting this is based on the fact that distributions of tracks on control chips, left in air for long time, are often very nonuniform. That could be explained by assuming that some areas attract positive radon ions from air while other areas repel them. Keeping chips in distilled salty water, and using a polyethylene protection, reduces exposure of CR-39 to air (from many days to no more than about 20 minutes). How can this hurt? Following Richard protocol, our control CR-39 chips are kept in a bottle with unused electrolyte, during electrolysis. This also protects chips from air. 

Only three people joined me in The Curie Project (all three are from Phys-L, an Internet list for physics teachers). Do you agree that the result would be treated more seriously if there were at least one or two electrochemists, with impressive credentials, among us?  

The goal of the project is modest; it is to verify Oriani's results--nothing more at this stage. Suppose we confirm that excessive tracks are reproducible on demand. Then we might have a good chance of obtaining financial support, from NSF or DOE, for more sophisticated investigations, for example, with electronic detectors. The goal would be to identify particles responsible for observed tracks, to find parameters influencing the rate of emission, and to test specific predictions based on theoretical models.


8) CMNS researchers did not reply to my appeals.
Why not a single CMNS researcher joined me in The Curie Project? Most of them probably agree that participation of Ph.D. electrochemists would be very helpful, as far are credibility of results is concerned. But they prefer to work individually, on their own projects, rather than to work together to verify already-published results. That is why progress is so slow in this area.

9) A message sent to three cooperating teachers
Practical details are usually not described in published papers. But details might help three teachers who will perform experiments in The Curie Project. I assume they read Oriani’s paper. Here is what I wrote: “I just started Experiment #7. All goes smoothly. Here is a description of details. Perhaps this will be useful.

1) Performing experiments at home is convenient. This allows me to record the current avery several hours, without any effort. Each experiment now takes three days. But etching will be done at school; in the chemistry lab, under the hood. This is because etching solution, if spilled, is much more dangerous than the electrolyte.

2) I have several well labeled containers. They are:

a) Unused electrolyte (a plastic bottle)
b) Used electrolyte (a plastic bottle)
c) Unused electrolyte for control chips (a plastic bottle with a wide opening)
d) Salty water for unused chips (a beaker)
e) Salty water for used chips (another beaker)
f) Distilled water (also a beaker)

2) Initially all CR-39 chips are in (d). To perform an experiment I remove two chips from (d), rinse them in (f) and set up an experiment.

3) Here is a very important detail. I make a pencil contour of each chip. Control chip is at once placed into (c); the chip is totally immersed.
Each experimental chip has a label, scratched on its surface. My convention is to position the chip in such a way that the scratched label is facing down (away from the electrolyte). Using Richard's terminology, I can say that my front surfaces are unlabeled while my rear surfaces are labeled. By careful; you must know which side is facing what, during the electrolysis. It is easy to make a mistake. I double-check the chip orientation after each experiment. Recognizing the side where the label is might be difficult. But it becomes easy after a right kind of illumination, and the angle of observation, are   found. In case of doubt; one can always do this under the microscope.

4) My cell is not as tall as Richard's. But I have enough of electrolyte above the anode to last three days. At the end of day 3 the level of electrolyte is about 8 mm above the anode.  I would certainly add a little of water if I noticed a possibility that the level might become much lower during the night, or during a prolonged absence. During the electrolysis, the current increases slowly from the initial value of 65 mA to about 90 mA. Following Richard's advice, I am not trying to keep the current constant (which can be done by changing the voltage or by adding distilled water to the cell).

5) I am using very light cables between the power supply and electrodes. Each cable is scotch-taped to a supporting stand, near the cell.  In that way the weight of electric cables does not contribute much to the weight of the column inserted into the cell. I practiced this  with during the first experiment (when a layer of polyethylene was used instead of CR-39 and when distilled water was used instead of the electrolyte).

6) At the end of the third day, I reduce the voltage to zero and turn the power supply off. The cable clips are disconnected from the Pt and Ni wires and electrodes are slowly removed from the cell. Then the used  electrolyte is placed into (b) and the cell is carefully dismounted. I examine the Mylar and write down what I see (one time I had a hole in Mylar). I also examine the CR-39 surface and write down what I see. (Is it dry or is it wet? It is always dry, but this should be checked.) During that examination I always see where the 0-ring was, because it makes a visible imprint on the chip surface. This allows me to draw a circle on the contour I made in the logbook (see item 3 above). This circle defines the area in which tracks will be counted, after etching. It is important to be organized, and to have a logbook. After drawing the circle, I rinse the post-electrolysis chip in (f) and place it into (e). The control chip is then removed from (c), rinsed in (f), and also placed into (e). Contours of chips, and scratched labels, will allow me to identify chips later.

7) Then I proceed toward the next experiment, etc., etc. Etching will be performed after all experiments are performed. Meanwhile all used chips are in salty water. I add some water to this beaker occasionally, to compensate for evaporation.

8) Let me know when you are ready (after observing tracks on the little chip that was irradiated with alpha particles, and after building a cell). I will send you enough chips for three experiments (two chips per experiment). I suggest we share the results on the same day; probably before the end of September. In this way we will not influence each other.”

10) Another message posted on CMNS list (7/4/2009)
I do not know why no one replies to messages I keep posting about The Curie Project. Perhaps no one reads them; perhaps CMNS researchers think that the topic is not worth discussing. Below is another message; will it generate some interest? This remains to be seen.

‘What does the adjective “reproducible-on-demand” mean? Strictly speaking, reproducibility is not possible because every measurement is associated with a random error. The same can be asked about the “highly reproducible,” about “reproducible,” and about “more or less reproducible.” This question, particularly important in our field, is prompted by Richard’s ICCF14 report 


The abstract of that paper contains the following: “However, not every electrolysis experiment produced nuclear particles so that total reproducibility was not achieved. Therefore another experimental technique has been developed which has shown the generation of nuclear particles in each of twenty five consecutive electrolysis experiments using heavy or light water solutions of lithium salts.” Does this statement mean that Oriani-type experiments are “reproducible on demand” ?

The purpose of The Curie Project is to verify that consecutive experiments demonstrate emission of nuclear particles due to electrolysis. Note that results from the first eleven experiments, in Richard’s Table II, were performed with the older CR-39 in which the mean density, on 16 control chips, was 26.4 tr/cm2 ( 12.1). The remaining 14 experiments were performed with the newer CR-39 in which the mean density, on 13 control chips, was 13.6 tr/cm2 ( 

Please help me to decide what kind of minimum results would be needed to conclude that The Curie Project validated reproducible generation of nuclear particles due to electrolysis. I will probably perform at least ten experiments. Three other participants will probably perform at least three experiments each. Richard's contribution will be in the form of results from experiments performed after ICCF14. Do you agree that a definition of "success" should be made before the results are known? 



11) Comments etc.

a) Ed Storms, the author of a recently publish CMNS book

posted this message on the CMNS list: ”I would like to extend the discussion that Ludwik has started. Two different types of experiments are possible. The first tries to demonstrate that a phenomenon is real.  Reproducibility is important when the reality of a phenomenon is in question. This the issue Ludwik   has raised with respect to the Curie project, i.e. how big must the result be before it can be considered real? The other kind of experiment accepts the reality of the effect, no matter how difficult it is to reproduce or how big it might be, and tries to understand how and why the phenomenon occurs. I think this is the kind of experiment we now need because the phenomenon has already been shown to be real   by a huge data set.

We now need to understand the variables and conditions that affect the   phenomenon. This means that people need to explore different conditions, not simply duplicate conditions that have been successful in the past. A very large parameter space awaits exploration, which requires trying many conditions, most of which will give nothing. However, the conditions that give a positive result become important   because they provide a boundary to the active parameter space. Experiments that attempt to exactly duplicate previously successful conditions provide very little new information other than to show that the collection of conditions created at the time, most of which are unknown and unmeasured, happen to combine to give a positive result. The only way to determine which of these conditions is important is to change them and see what happens. For example, suppose the radiation is very sensitive to the concentration of PdCl2 and the chosen value happens to be on the edge of the critical value. One person happens to be on the high side of the designated value and gets a positive result and another person happens to be on the low side and gets no   radiation. Without knowing the true effect of the PdCl2 concentration, the result will provide no useful information other than to give one more positive result to show skeptics.  This is only one example of hundreds of possible variables that might have an effect. We need to start the long and boring exploration of these possibilities rather than trying to simply reproduce the effect.  Without a useful theory (perhaps too many theories) to guide this search, the work needs to be done mainly by trial and error methods until some of the important variables have been identified.

On the other hand, I agree that discovery of a reproducible method would make such an exploration of variables easier. For example, it is easier to find first base if the location of the ball field is known in advance. However, we don't even know the city in which the ball field is located. We need to send out explorers to every city we can   think of to locate the ball field before we start asking the location of first base.  We have shown by chance that the ball field actually exists, but the map was never properly drawn by the successful explorers. We don't need to once again prove the ball field exists. We only need to locate it in a way that results in a useful map.

This means the variables we have available in the experiments need to be changed by amounts that are easy to measure in order to be sure that the variables are not affecting the result by small chance variations. This takes time so, I suggest, the sooner we get started the better.”

Hmm, “people need to explore different conditions, not simply duplicate conditions that have been successful in the past.” That might be a good explanation CMNS researchers do not want to participate in The Curie Project. But shortly then I received a message from a after young researcher wrote does want to collaborate. Announcing that event, I wrote:

“1) I am happy to report that one researcher from our list is also going to participate in The Curie Project. That is very welcome; the more the better.

2) That person wrote: "My goal of working on the Curie Project is so that I can give simple reproducible experiment to  scientists who have venture capitalist connections and let them run it and get them excited about it. The key is that I need something relatively simple and highly reproducible.

I have some CR-39 and I'm gathering the equipment over the next few weeks.  Do you have any useful components that I can buy from you?  In particular, where to buy the Mylar?"

3) We have practically no 6-micron Mylar left. Can someone send me some, to be sent to project participants? Each Oriani-type experiment uses about one square inch of Mylar. Five participants, performing 5 experiments each, would use about 30 square inches. I would be happy  to pay for about 300 square inches, in anticipation of future experiments.

4) Will the "reproducibility on demand" be confirmed in The Curie Project? We will probably have a clear answer in September.

5) Please help me to decide what kind of minimum results would be needed to conclude that generation of nuclear particles due to electrolysis is "reproducible on demand." I agree with Ed that a lot of investigations will be needed, after reproducibility on demand is accepted. Each sequence of several experiments would begin with an experiment that always produces excessive tracks. Other experiments in that series will be conducted to answer a simple question--what effect a change in one parameter has on the mean track density? A change in one parameter could be:

a) Replacing Ni cathode by a Pd or Pt cathode.
b) Placing the cell into a magnetic field.
c) Replacing one kind of electrolyte by another.

6) I also agree with Ed that this kind of investigations would benefit enormously from specific theoretical (computational) predictions. Something like this:

a) My theory predicts that replacing Ni by Pd would increase the mean track density by at least one order of magnitude.
b) My theory predicts that magnetic field will reduce the mean track density by a factor of about two or three
c) My theory is that using PdCL2, instead of Li2SO4 will increase the mean track density by two orders of magnitude.

Theoretical predictions should be specific; the more specific they are the more useful they are likely to be.”

In a private message one CMNS researcher wrote:

“ 1) In my opinion, in addition to reproducibility one might test for additivity (number of tracks of particles due to electrolysis should be proportional to duration of electrolysis). And it should be proportional to excess heat. Naturally, this should be true when other conditions are strictly identical. My prediction is that there will be no proportionality; there will be some kind of saturation. 

2) In applying magnetic field, one must not forget about the terrestrial field. Do not be disturbed by its small size. In the final analysis, we do not know what is the nature of terrestrial magnetic field, no matter what the experts say. By the way, terrestrial field has two components. That is why it would be useful to apply magnetic field in both directions. My prediction is that the form of tracks might change radically when the vertical field (exceeding 20 Gs) is applied. 

3) S-based electrolytes might produce stronger effects than Cl-based electrolytes. 

Ludwik, I am aware that these comments are useless. There is no need to answer, or to pay attention to them. “

This is not what a useful theoretical prediction; the author is aware of this. Why is S more desirable than Cl? Why should vertical magnetic field produce the described effect? What is the cause of the described saturation? Predictions without justifications are not likely to be productive. Cooperation between theoretical and experimental scientists is necessary. Theoretically inclined people should be informed about essential experimental details; experimentalists should be informed about assumptions on which theoretical calculations are based, and, if possible, with mathematical equations as well.

12) More comments
In a message posted on CMNS list, another scientist wrote that 100% reproducibility is not a necessary requirement for the existence of a phenomenon. Some CMNS effects are not reproducible on demand because we do not know how to control them.
Responding to this I wrote: “That is true. But, according to Table II in

each of the 25 consecutive experiments “has shown the generation of nuclear
particles.” That is why the initial goal of The Curie Project is to verify the high level of reproducibility. Suppose we confirm that Richard’s protocol yields highly reproducible evidence of chemically-induced emission of nuclear projectiles. Suppose that a paper, coauthored by all project participants, is submitted to Physical Review, or to another journal that rejected Richard’s original manuscript. Suppose that this fact is widely publicized (with help of honest TV journalists, etc.). I suspect that this might produce at least one desirable result--publication of our paper. That would be an important step toward the end of unfair discrimination.

At the other extreme we might find the opposite situation--no one but Richard observed a clear signature of a nuclear process due to electrolysis. That would be the end of the story, as far as high level of reproducibility is concerned, at least for me. And what if some of us, excluding Richard, observed the clear signature while others fail to observe it? That would be a situation described by Brian. We would say that Oriani effect is not very different from many other CMNS effects--it seems to be real but not yet reproducible on demand. Of course, we hope that the exceptionally high level of reproducibility will be confirmed. That would be our “success.” But we should be prepared for different outcomes--a “failure” or a “partial success.”

Another message posted by Ed Storms
Responding to an interesting observation made by another CMNS researcher (also about importance of reproducibility) Storms wrote: “I agree with what you say and with the logic behind Ludwik's approach. However, I do not agree that a reproducible way to make pits in CR-39 will convince skeptics that LENR is real. First of all, the measured flux is very small and is not clearly related to the other kinds of evidence for a LENR effect.  Second, the method is not well understood by the scientific establishment, which would be necessary for the effect to be considered proof. The neutrons claimed by Pam et al. get deserved attention for reasons that Ludwik and other successful studies cannot duplicate. The SPAWAR results were examined at a respected laboratory using high level professional equipment. Even if a procedure is found allowing anyone to make the required pits, the results will not prove that important heat can result from such a reaction. Such a procedure will show only that low level radiation can be produced, which I agree is important but not important enough to interest serious founders, especially ones who want to make energy. Cold fusion is important because it might make useful energy and it implies a nuclear reaction rate far in excess of the SPAWAR results, which is in excess of what can be explained by any theory.

Instead of trying to simply reproduce one small part of the effect this way, I suggest the best approach is to focus on trying to understand the effect, either what is required to make it happen at a higher rate or the nature of the radiation. Granted, getting the effect to work is essential before it can be explored. Nevertheless, I'm suggesting the emphases be placed not on reproduction but on understanding, with reproduction being only the first and not even the most important step. The stated goal would not be to provide a method people could use in their own lab to prove to themselves that the effect is real. Instead, the goal would be to use the reproducible procedure to discover something important about the effect. While this would happen automatically when a reproducible procedure is found, stating it up front with emphasis gives the effort more importance than simply trying to achieve reproducibility.”

And here is the message posted by me: “Brian, Ed and Tom clearly identified two distinct aspects of our dilemma: we need to understand what is going on, and we need protocols yielding reproducible CMNS signatures.

Why do I think that focusing on reproducibility is now more important than focusing on understanding? Because, as stated above, we do not have an "accepted baseline case from which controlled explorations can be made." Suppose I want to learn about the effect of X on excess heat, or on the rate of a transmutation. I perform ten experiments and results scatter from nothing to a lot. What only thing I learn is that experiments do not yield information about the effect of X on the selected effect. This can be contrasted with a situation in which results from consecutive experiments are more or less similar. I obtain the mean value of the result and start another sequence, for example, after increasing or decreasing X. This would tell me something definitive about the effect of parameter X. Focusing on understanding is not possible without a reproducible baseline.”

On July 7, 2009 I wrote:
How can one objectively distinguish real excess tracks from apparent excess tracks? I had a plan for this. Unfortunately, the number of participants in The Curie Project is too small to implement it. (Not counting Richard, whose results we will be verifying, there are five of us: two high school teachers, two engineers--who are on this CMNS list--and myself.) Suppose the number of participants is 25, instead of 5. Suppose that each of us performs 6 experiments. This would produce 25*6=150 measurements of track densities. Let me ignore two trivial outcomes-- (1) every measurement is positive (confirming that track densities due to electrolysis are significantly higher than on control chips), or (2) every measurement is negative.

The third possible outcome is that some measurements yield positive results while others yield negative results. I have a plan for this difficult scenario (see below). Do I explain it clearly? Would it be acceptable to most scientists? What is wrong in this plan? What is a better alternative?

a) Each participant has a series of six results. This gives us 25 mean track densities, one from each participant.

b) A histogram of mean values is constructed. Suppose the first bin is for means  between 0 and 30, the next is for means between 30 and 60, etc.

c) Yes, 25 is not a large number. But it might be sufficient to answer a simple question--is the distribution consistent with the   bell- shaped curve?

d) Reasonable consistency would then be interpreted as an indication that excess tracks are real (the mean of means becomes our estimate of the “true” excess track density). Clear inconsistency, on the other hand, would prevent us from making this assertion. Yes, I would also prefer 250 participants, instead of 25. But this is totally   unrealistic, in our situation. Suppose 12 means are in the first bin,  6 means are in the second, and the remaining means are nearly randomly distributed (a tail) among bin 3, bin 4, bin 5, etc. That would certainly not be consistent with the expected bell-shaped   distribution.

The rational for depending on the histogram of mean values is based on the so-called “central limit theorem” of statistics.  If occasionally- observed excess tracks are real then the distribution of mean values must be Gaussian, provided the number of mean values is very large. Existence of a “true value” is equivalent to the existence of a large unchangeable population from which individual samples (yielding mean values) are randomly selected.”

Ed Storms responded:
“Ludwik, I think you are making this problem too difficult.  If you do   not see significantly more tracks on the CR-39 that is near the   cathode compared to locations more distant, then you have failed to   demonstrate reproducibility. The exact number of pits does not   matter.  Your object is to demonstrate the reality of the effect, if   the reality cannot be recognized by eye without complicated   mathematical analysis, it has failed. The Spawar results were easy to   recognize as localized radiation simply by visual inspection of the   CR-39. You need a similar result if you intend to claim to reproduce   this result.   Of course, complicated analysis is required to identify   the nature of the radiation.  However, this is a secondary result that   does not appear to be part of your procedure.  On the other hand, if   your object is to study the effect and show which variables are   important, as I suggested, then you need to make qualitative   measurements including when pit density is low, because you need to   relate pit density to the different conditions being explored.    However, this does not appear to be part of your procedure either.”

My reply:
“1) Yes, situations in which statistical analysis is not necessary are preferable.
2) Particles producing "tracks due to electrolysis" might be produced in the electrolyte, as reported at  ICCF10. The first thing I will do, if reality of excess tracks is confirmed, will be to change the distance between the cathode and the chip from nearly nothing to ~2 mm (which is more than sufficient to stop alpha particles emitted from the cathode). The purpose of The Curie Project is nothing more than to get a clear yes-or-no answer about reality of excess tracks via Oriani's protocol. Investigations of parameters is tempting; but we should not be distracted. This would be the next step.”

On July 8 I posted this message:
“ 1) I hope some of those who are not able to perform The Curie Project experiments will be able to participate in the analysis of results. Our data will probably be available in September. But this should not prevent us from discussing strategies. Please save this message; it might be useful later.

2) As you know, row data are summarized in Table II of Richard's ICCF14 paper

In table I of this paper he tells us that the first 11 experiments were performed using CR-39 where mean background was 26.4 tr/cm2 (stdev=12.1 tr/cm2). Better CR-39 was used in the remaining experiments (mean=13.6, stdev=6.8)

2) What do we want to accomplish? We want to answer two questions: (a) are our results statistically consistent with Richard's results? (b) Do our results (and Richard results, summarized below) support the claim that his protocol yields a highly reproducible signature of a nuclear effect due to electrolysis?

3) The most interesting Richar'd result comes from his experiment 9 (too many tracks to measure). But I am excluding it; I am going to use only what was actually measured. The first two experiments (SPAWAR-type electrolysis) are also excluded. In other words, we have 21 experiments in which track densities were measured. Each experiment provided two numbers (after first etching)--track density on the front surface and track density on the rear surface. For the time being I am place all measured densities into one set of 21*2=42 results. What follows are measured track densities, listed in ascending order; ignore the number in parentheses; it identifies the experiment, not the number of tracks per square centimeter (like the first number in each line).  


4) The mean value is 122 tr/cm2; the standard deviation is 124 tr/cm2, and the median is 73 tr/cm2 (in other words, one half of the results produced 73 tr/cm2 or more). A large difference between the mean and the median indicates that the distribution is strongly skewed.  This is confirmed by   following histogram:

    bin 1 (0-30 tr/cm^2) <------------6 results
   bin 2 (30-60 tr/cm^2) <---------------------- 10 results
   bin 3 (60-90 tr/cm^2) <------------------ 9 results
  bin 4 (90-120 tr/cm^2) <-------- 4 results
 bin 5 (120-150 tr/cm^2) <----2 results
 bin 6 (150-180 tr/cm^2) <--1 results
 bin 7 (180-210 tr/cm^2) <------3  results
 bin 8 (210-240 tr/cm^2) <--1 results
 bin 9 (240-270 tr/cm^2) <0 results
bin 10 (270-300 tr/cm^2) <--1  results
bin 11 (300-330 tr/cm^2) <0 results
bin 12 (330-360 tr/cm^2) <----2 results
bin 13 (360-390 tr/cm^2) <0  results
bin 14 (390-420 tr/cm^2) <--1 results
bin 15 (420-450 tr/cm^2) <--1 results
bin 16 (450-480 tr/cm^2) <0 results
bin 17 (480-510 tr/cm^2) <--1 results

5) How to deal with a situation in which the distribution is not Gaussian, and when the standard deviation is about the same as the mean value? These are real experimental data. Knowing that the standard deviation is 124, what is the level of confidence in the following statement "The mean=122 tr/cm2 is significantly larger than the background." The level of confidence would higher than 90% if the distribution was roughly bell-shaped and if the standard deviation were something like 40 tr/cm2, or less. ”

Ed Storms responded:
‘If the standard deviation is as large as the value, my suggestion is to ignore the results and work to improve the signal rather than applying math. No one will believe the results no matter how you "skew" the data.’

My reply July 9, 2009
“I have no doubt, at the intuitive level, that Oriani's results provide a very strong evidence that a difference between the mean track density--on his 42 post-electrolysis CR-39 surfaces  (122) and the mean track density on his control chips (26)--is highly significant. But I do not know how to calculate the level of confidence for this statement. That is not a good reason for ignoring the results.

Yes, a significant difference does not guarantee that others will report similar results; that is what the Curie Project is about. The more participants we have the better. IT IS NOT TOO LATE TO BECOME A PARTICIPANT.

Intuitive evidence is  not a guess; it must also be justified. Here is the justification for my use of the term "intuitive level" above:

Richard tells us that experiments from #4 to #11 were made with CR-39 where mean background control density was 26 tr/cm2. This amounts to 8 experiments, or 16 measurements of track densities. Among these I see only one result where the density was less than 26. All other track densities, including the four from the very impressive experiments #5 and #11, are larger than 39. Please let me know why an honest skeptic would reject this argument.

The last 14 experiments were made with CR-39 where the mean background control density was 14 tr/cm2. This amounts to 28 measurements of densities. Among these I see only three results where the density was less than 14. All other densities are larger than 25 tr/cm2, including ten densities higher than 100 tr/cm2. Please let me know why an honest skeptic would reject this second argument.

How can anybody doubt that the difference between Richard's control chips results and his results from electrolysis experiments are dramatically different?

Is it not true that  Richard's evidence (that post-electrolysis densities are higher than control densities) is very satisfactory? “

My message posted on July 10, 2009
“Some phenomena are reproducible but rare. In other words, one does not always observe them. How to test for reproducibility of such results? By comparing what happens with what should happen if results are reproducible (reproducible results are expected to to satisfy the Poisson distribution should). The shape of that discrete distribution depends on only one parameter L (lamda). Both the mean value and the standard deviation are equal to L. Calculations are no longer tedious and error prone; use the online calculator at:

Suppose a reproducible event is claimed to occur only twice per day (on the average). I attempted to detect it in ten consecutive days nothing was found. One does not need to be very sophisticated to conclude that the claim was not validated by me. But this can also be demonstrated by using the Poisson distribution. What is the expected (most probably) number of observations in 10 days? it is 10*2=20. What is the probability to observe 5 observations in ten days? I make L=20 (most probable result) and x=5. According to the Poisson distribution (use the calculator above) the probability for X=5 is 0.000005. And the probability that X=0 is practically zero.

But in some cases intuitive predictions are not obvious. Suppose 16 observations were made. Is this consistent with the expected value of 20? For X=14, the probability is 0.039. This is not much smaller than 0.089, the probability for X=20.  In other words observing something only 14 times in ten days, instead of 20, is not very unlikely.

Let me apply this approach to Richard's observations.  According to my summary (posted two days ago) based on Table II in

Richard examined 42 CR-39 surfaces and found 15 cases in which densities exceeded 100 tr/cm2. According to Poisson's distribution (L=15 and X=15), the probability of observing exactly the same result as Richard is 0.102. Another researcher also measures track densities on 42 surfaces and finds that track densities exceeding 100 are on only 12 surfaces. Is this consistent with Richard's observations? The probability of that outcome (L=15, X=12) is 0.083. That is not much smaller than 0.102, the probability of the most likely result. I would say that 12 is in good agreement with Richard's result, 15. But what about X=5? In that case the probability would be 0.002. I would say that Richard's result, 15, is not reproducible.

Is there anything wrong with this approach? It is highly desirable to have an agreed upon test for reproducibility before The Curie Project results are known. The same approach, by the way, can probably be used for other skewed CMNS distributions. Excess heat, for example, as summarized in Ed's book,  was more often observed at low powers than at high powers. That is also a very skewed distribution.”

Ed Storm responded:
“You have two questions to address, Ludwik. Do the pits result from a random process or are they the result of rare special conditions? Your approach assumes the former, which is not very interesting. You are trying to show that a rare process is initiated under certain conditions. This is not a random process, hence the math you use does not apply.  You need to show that the effect is larger than would be expected from a prosaic process.  You need to identify the likely prosaic processes and show that they do not occur at the observed level. This requires changing the conditions in a known way and seeing the effect of these changes.  Simple duplication to which math is applied based on a random process means nothing.”

My posted question (after quoting the first question above):
“What kind of random process would produce many more tracks on experimental chips than on control chips? Richard's control chips are treated in exactly the same way as experimental chips, except for one thing. Control chips are in the bottle with unused electrolyte when experimental chips are exposed to a thin Mylar film supporting the cathode.”

P.S. (after several hours of silence from the list). What preoccupies me does not preoccupy others. Each researcher has his own priorities. But researchers do not share their experimental problems with each other very often, except at formal meetings. That is rather unfortunate. :-)

I am deliberately trying to separate two questions: (a) How to decide, that The Curie Project results (when available) are consistent with Richard’s results, and (b) whether or not a given set of results provides evidence of a nuclear process due to electrolysis. The second question is more interesting and more important. But, for the time being, I am focusing on the first question. This should not prevent us from addressing the second question, in view of what was reported by Richard in

Our sets of results might be consistent with each other without providing evidence for a new nuclear process due to electrolysis. Another possibility is that our sets of results are not consistent with each other but each provides evidence of the desired CMNS process. What is wrong with trying to answer the first question before addressing the second one?

Please help me to find an agreed-upon method for objectively answering the first question. Richard results were presented at ICCF14; The Curie Project results will probably be available in September. It would be useful to agree on the method before the results are known. Should we count cases where the mean density exceeds 100 or should we count cases where it exceeds 60 tr/cm2 ? How to objectively justify the choice?

In principle, our two sets of results should be consistent (or inconsistent) no matter what lower limit is chosen. But our sets suffer from small numbers of measurements. What is true for a large set of measurements, such as 400, is not necessarily true for a set of only 40 measurements. Here is one possible choice (upper limit 60 tr/cm2) and its justification:
a) 16 control chips (older CR-39) showed the mean of 26.4 tr/cm2 (stdev=12.1).
b) 13 control chips (newer CR-39) showed the mean of 13.6 tr/cm2 (stdev=6.8).
c) The weighted average is very close to 17 tr/cm2. We chose 60 because it is more than three times larger than 17. (Also note that 26+12+14+7=59).”

On July 11, 2009, at 9:33 AM, Edmund Storms wrote:

“Ludwik, although the electrolyte does not attack CR-39 before electrolysis, the situation can change when electrolysis starts. Many very reactive molecules are made during electrolysis and I expect some will pit CR-39. This pitting will be tricky to separate from radiation because pitting, like radiation, will be less the further the distance from the cathode. In addition, a protective covering designed to stop chemical attack will also partially stop radiation. Of course, a careful study can separate these effects.  Consequently, some of the pits could be caused by chemical attack and their number will appear to be random depending on a combination of many uncontrolled conditions. Since chemical attack is the most likely source of pits, the experiment needs to explore conditions that can clearly separate this possibility from the effect of particles.  When very few pits are seen, this becomes a difficult problem.  Because CR-39 accumulates any effect, the rate can be very small, hence might be caused by any number of prosaic processes, some of which might be hard to predict without careful study. In contrast, the Spawar study produced so many pits that many of these effects could be eliminated. The Curie Project needs to see a similar large number of pits.”
justify the

My reply: “Thanks Ed,
1) I am neither a chemist nor a material scientist. But I will be very interested in what other electrochemists are going to say. In some Oriani's experiments control chips were actually in the cell with the Mylar, the cathode and the electrolyte above. The only difference was that the current was zero, rather than many milliamps. In other words, the effect is due to the electric current. You are probably thinking about chemical reactions caused by the current. Right? The electric wattage is very low (about 0.6 W in my case) and the cathode is in thermal contact with the cold electrolyte.

2) Are you aware that according to the most recent article, published by the SPAWAR team, about 90% of pits disappear when the chip is protected from the electrolyte by a thin layer of Mylar?  Yes, the remaining 10% is still a lot of pits; probably "too many to count." 

3) The codeposition procedure (covering chip #7 with Mylar), used by F. Tranzella, on the other hand, yielded a very different result, as  reported in Catania (Lipson et al., page 182 of our proceedings). The track density became comparable to what is typically observed by Oriani. Furthermore, Tanzella's tracks were identified as protons of several MeV. What should we think about this apparent controversy? Would you agree that your suspicion of chemical effects is also applicable to Tanzella's results?  If not then why not?

4) Two days ago you asked for the tiny bubbles in my cell. The electrolyte becomes "milky" about two or three seconds after the current is turned on. But it remains "milky" for about six to eight seconds, after the current is turned off.”

Ed’s reply:

“1) The current creates many unstable chemical species that can react with relatively inert material during their short lifetime.  For example, stainless steel is stable in H2O+LiOH but is not stable when the fluid is electrolyzed. These unstable, reactive compounds are made at the electrodes which result in a concentration that is greatest near the electrodes.  Therefore, the drop off in pit density away from the cathode can be caused by how far the reactive material can diffuse before it decomposes rather than by absorption of particles in a greater length of fluid.

2) Thanks, this is important information. The question is whether the Mylar is fully protective or whether some of the reactive species can eat through the Mylar. However, the pits on the opposite side that is not exposed to the electrolyte would seem to be caused by real energetic particles.  Also the pits that show clear characteristics of particle effects also provide good evidence for particle emission. As long as a large number of such pits are present, I would accept the conclusions.

3) I agree, possible chemical effects become more important when the pit density is low.

4) Interesting.  You are obviously making many very small bubbles that seem to be related to the nature of the cathode surface.”

A message posted by Scott Little (7/12/2009)

“Ludwik, One of the problems you face in interpreting results from your Curie project is this: Do the pits look right? IMHO, this subjectivity is one of the greatest shortcomings of CR-39. Please look at some of the photos we posted in our SPAWAR report here:
Alpha particle pits are fairly uniform in size and have a random spatial distribution over the surface of the CR-39.  As a result of this distribution, overlaps are relatively rare.   Compare this with what we called the "soap bubble" appearance which we are virtually certain is the result of chemical attack on the CR-39.
Note also the appearance of pits caused by various mechanical damage means.  Again there is a high number of contiguous groups of pits, which is not expected from energetic particles. Yes, I understand that there is a hypothesis for multiple tracks from energetic particles but surely the overall cross-section (nuclear cross-section multiplied by a geometric cross-section expressing the probability that the event is properly located/oriented so that it develops during etching) for that interaction is very small compared to the prevailing interaction that produces a single pit.
If you find lots of pits in your Curie experiments, please post lots of good microscope photos so that all of us can help you judge whether or not they look right.....:)”

Ludwik’s reply:
“Thanks Scott,
1) Pictures of our pits will be shown at my website. And I will be happy to send etched chips to those who might be interested. The goal of The Curie Project is to verify Richard's results. His pits are very similar to those produced by alpha particles. That is why only such pits will be counted. Will we also observe pits (on both sides of CR-39), at densities exceeding what is typical for his control chips (mean about 20 and of about 10 tr/cm2)? That remains to be seen.

2) Referring to Oriani-type experiments, you wrote: "We performed background counts on both sides of 8 chips. The average was approximately 100 pits/cm2." That amounts to 16 surfaces. I suppose your four-years-old CR-39 chips were from the sheet whose part you sent me four years ago. Please confirm this. My recent measurements show that the mean background, for that sheet, is 26 tr/cm2. I plan to verify this. If confirmed then your average of 100 tr/cm2 would be consistent with Oriani's measurements, at least qualitatively.”

Ludwik (replying to Ed’s message; 7/13/2009)

“Dear all, 
1) My suggestion is to start using electronic detectors instead of CR-39, as mentioned in an earlier message. Placed in air, next to a Mylar window, they would not be sensitive to "unstable chemical species." The same, by the way, would be true for a CR-39 detector placed about 1 mm away from the Mylar window. Furthermore, as many of you know, silicon-surface-barrier detectors do much more than to accumulate signals produced by nuclear projectiles; they provide information about their energies and about times of emission (does it occur in bursts? does it change when the electric current is increased? What happens when the current is turned off? What happens when the magnetic field is applied? What happens when the electric current is redirected to another cathode, when several drops of something is added to the electrolyte, when . . . ? etc. etc.).

 2) Why should several hundred tr/cm2 be called "very few pits" when Oriani's mean control chips density is about 20 tr/cm2 (and the stdev is about 12 tr/cm2)?

3) The problem with artifacts is more general. Three kinds of  CMNS phenomena: "excess heat," transmutation," and "emission of radiation" were already named in 1989/1990. I am thinking about Bockris, Fleischmann, Jones, and Pons. Two objections were immediately raised by skeptics: (a) results are not reproducible. (b) results might be due to artifacts. How do you know that your excess heat is due to a nuclear process, and not to another (known or unknown) phenomenon? How do you know that the so-called "new elements" were not already present in your materials (or introduced during manipulations)? How do you know that CR-39 pits are not due to "reactive molecules," as suspected by Ed?   

4) The immediate goal of The Curie Project is to verify that Oriani effect is reproducible. This is much easier than to provide convincing evidence against artifacts. Why is it so? Because the number of conceivable artifacts is practically unlimited. Knowledgeable people can always invent reasonable artifact scenarios. A claim for a discovery should be tentatively accepted after satisfactory arguments were presented against two or three most obvious artifacts. That is what usually happens in most areas of science, as far as I know. It is practically impossible to answer all conceivable objections when a discovery is first announced. We do not except a newborn baby to walk or run; this comes later. Tentative acceptance is a beginning of more refined investigations. Some initial benefit of doubt is usually given to qualified researchers. Why shouldn't we tentatively accept Oriani's effect as real? Convincing evidence against obvious artifacts has already been presented by both Richard and the SPAWAR team.  

5) Yes, I know that reproducibility is a precondition for all debates about experimental data. A negative result is not convincing unless experiments are known to be reproducible. Reproducibility is mother of science. 

6) SPAWAR results, as far as I know, are reproducible when CR-39 chips are exposed to electrolyte, as in The Galileo Project protocol. But they are not reproducible when Mylar is used. This observation is based on what was known two years ago (Catania workshop) and what was published last April (by the SPAWAR team). Why was the controversy not mentioned in their last paper? What is the present status of reproducibility of SPAWAR results? ”

Appended on July 17, 2009
I do not know why numerous attempts to obtain information about the status of the Tanzella-SPAWAR discrepancy did not produce any new information on our CMNS list. Authors of papers to which I was referring are the list subscribers. Something important is probably behind this strange situation; what is it? Giving up on this, I posted another message. Last night I wrote:

“Theoretical topics, such as HUP, QM, QCD, standard model, existence of energy states below the ground state, etc., are certainly interesting to those who are qualified. But where should serious debates on such topics take place? Michal Gryzinski, for example,  debated his ideas at gatherings of theoretical physicists well equipped to understand him, to criticize him, to help him, to explain his ideas to others, etc. Is the CMNS field appropriate for debating such issues? If I had something to contribute in the field of "quantum chromo-dynamics," for example, I would address those who know this theory well, and who use it routinely.  I would not address a conference (or list)  for geologists or chemists. Yes, QCD might eventually help them; but that would not be my excuse for addressing them, instead of addressing theoretical physicists. I am afraid that only a small fraction of us is equipped to discuss advanced theories.

Our field is already very controversial. We were not able to convince mainstream scientists that our claims (excess heat, transmutations, and nuclear projectiles) are valid. In order to succeed, we should stay away from additional controversies, at least for the time being. I am afraid that  embracing other controversies could hurt us at this time. Do you agree?

The interview with Fleischmann, published today, reminded me of another case where a premature attempt to explain experimental facts backfired. Suppose the claim made by Fleischmann and Pons were limited to  "we discovered unexplained heat." This would lead to arguments about experimental techniques, the field in which Martin was a recognized first class international authority. But he said "therefore it must be nuclear," as quoted in the interview. Undesirable consequences of this premature statement are still with us. It would be better if they waited for independent confirmation of excess heat before speculating about its origin. Linking an interesting experimental fact with a controversial theoretical topic  is an open invitation for troubles, with referees and with other scientists.  Do you agree?”

A nearly-immediate reply, from Ed Storms, was about importance of understanding. How can I disagree with this? Science is not only a collection of experimental results; it is also a set of explanations. According to American] “National Science Teachers Association”

“A primary goal of science is the formation of theories and laws, which are terms with very specific meanings.

*) Laws are generalizations or universal relationships related to the way that some aspect of the natural world behaves under certain conditions.

*) Theories are inferred explanations of some aspect of the natural world. Theories do not become laws even with additional evidence; they explain laws. However, not all scientific laws have accompanying explanatory theories.

*) Well-established laws and theories must be internally consistent and compatible with the best available evidence; be successfully tested against a wide range of applicable phenomena and evidence; possess appropriately broad and demonstrable effectiveness in further research. . . . While science and technology do impact each other, basic scientific research is not directly concerned with practical outcomes, but rather with gaining an understanding of the natural world for its own sake.”

That answer to “what is science?” has not changed substantially since the times of Gallileo and Lavoisier. Science was, and continues to be, the basis for our technological civilization.

A discovery can be made experimentally or it can be made theoretically. Here is one well-known example. A Danish astronomer Tycho Brahe studied motion of planets and recorded their positions. His assistant, a German scientist Johanes Kepler, analyzed the result and formulated lows of planetary motion. His first law stated that trajectories of all planets are elliptical. His second and third laws were about the dependence of velocities of planets on their distances from the sun. Results of Brahe's investigations were raw experimental facts about individual planets; Kepler's law were generalization based on these facts.

Knowing about Kepler's laws, Isaac Newton discovered a theoretical explanation of planetary motion. He showed that Kepler's laws can be mathematically derived from only one assumption. That assumption is the existence of an attractive force between all objects in the universe. To be consistent with Kepler's laws, this gravitational force must be directly proportional to masses and inversely proportional to distances between the objects. Newton's theory was an important contribution to our understanding of nature; it is now used to design man-made satellites. Progress of science is based on experimental and theoretical investigations; they go hand and hand with each other, in a long run.

But neither Ed, nor other who responded, addressed the issue of negative effects of premature explanations. Instead of focusing on replications, readers (and referees) are likely to focus on explanations. Experimental data and theories go hand in hand but, in the final analysis, theories which disagree with facts are rejected, not the other way around. Ed Storms wrote:

“The issues involving QED, QM, and the Mills approach are basic to understanding LENR.  Any theory must be consistent with what has been observed outside of CF, which includes phenomenon related to and explained by QM and the other models. Once a nuclear reaction takes placed in a CF environment, the consequence must be consistent with what has been observed to happen when similar nuclear reactions occur in other environments.  In short, we need to see a link between CF and other observations.  This means the theories used to explain these observations need to be understood before they can be applied to CF in order to show a relationship to other phenomenon.

Clearly QM is not fully understood by conventional science even when it is applied to accepted observations, hence the debates.  In the process of understanding CF, I expect an understanding of the basic ideas behind QM will increase. I'm sorry to say we need to understand these ideas whether we like it or not.  Also, we should not compromise our discussion just because some scientists are closed minded even about subjects that do not involve CF.  These narrow minded people are clearly not able to evaluate reality even in normal science much less when it comes to the challenge of CF.  We should not let their deficiencies influence our approach.”

Two people wrote that they enjoy theoretical discussions on the CMNS list. Responding to them, I wrote:
“ Me too. Yes, we do need theories to guide us. In the long run, experimental data and theories go hand and hand. Keep in mind, however, that a scientific theory conflicting with experimental data must be rejected, not the other way around (as some skeptics do). My main point was that readers and referees should not be distracted from experimental procedures by controversial theories. 

Let me bring another illustration. It is the ICCF11 paper of G. Lochak and L.Uretskoev. They described a procedure by which the isotopic composition of a Ti wire was changed. What was the purpose of trying to explain a reproducible laboratory result with magnetic monopoles? This was not a good strategy. My articles would have a different title, for example, "A distorted isotopic ratio in Ti: how can it be explained?" I would keep my theory on the back burner, waiting for additional replications, and for the acceptance of my experimental claim. Why? Because the topic of magnetic monopoles is known to be controversial. An explanation based on a non-controversial theory, on the other hand, would be worth mentioning. That would probably be helpful. 

Linking an interesting experimental fact with a controversial theoretical topic is an open invitation for troubles, with referees and with other scientists. That was the main point of my message.”

Ed Storm responded: “I agree with you, Ludwik. Experimental papers should not attempt an explanation unless the experiment was designed to answer a question theory raised. Speculation about an experimental result should be reserved for a paper devoted solely to that effort. Mixing theory and experiment invites trouble as you say. Unfortunately, people seem to be unable to restrain themselves in trying to explain rather than simply reporting.”

This prompted me to post a short observation: “Why are people unable to restrain themselves? Because we were trained to think that science is much more than factology. And this is true.” Elaborating on this Ed Storms wrote:

“The problem is in our genes. The human species has a need to explain. We can't help it. This allows great progress but also creates the conflict in all subjects because we take our explanations too seriously as the truth. In fact, the explanations are very seldom correct much less being the truth. Unfortunately, our species was also given an ego that has no humility. Consequently, most people will defend their chosen explanation to the death no matter how simple-minded the explanation might be. You can see this process operate most clearly in religion and politics. Science tries to overcome this problem with limited success.”

Appended on July 29, 2009
A CMNS researcher posted a message about hypothesizing. Responding to this message I wrote: “ (1) What is a scientific HYPOTHESIS ? It is neither a theory nor an   experiment. It is a preliminary ASSUMPTION made by a scientist. Some call it intuitive thinking, others refer to it as a speculation, educated guess, hand-waving, etc. (2) I think that linking our field with controversial assumptions, such as hydrinos, is likely to result in more harm than good (at this stage).”

Scientific method in physics consists of techniques to investigate natural phenomena. Does a particular reported phenomenon exist? What is known about that phenomenon? How to explain it in terms of truth already accepted by scientists? The first two questions are addressed by experimentalists; the third is addressed by theoreticians. Experimental data are validated in terms of reproducibility. Theories are validated not only on the basis of logical (mathematical) correctness but also on the basis of specific predictions verified in reproducible experiments. Some theories are closer to reality than others.

This website contains other cold fusion items.
Click to see the list of links