open access


Our objectives were to study the effects of PEMF on a non-animal living system with a simple and unambiguous biomarker: seed germination rate. We selected seeds representing three groups: high- intermediate- and low-germination-rate seeds (lettuce, bell pepper, and strawberry, respectively). PEMF was applied at 15 pulses per second (pps or Hz) to strawberry and lettuce seeds, and 3 or 15 pps to bell pepper seeds, using only one well-defined PEMF pulse waveform shape and intensity. The only PEMF variables between groups were therefore PEMF pulse rate and total number of pulses, which was taken to be the PEMF dose, equivalent to the duration of exposure at a known pulse rate. In the case of lettuce, we studied the effects of germination using only one PEMF dose compared to no-dose (PEMF -) controls, and no interactions with other pre-planting procedures. In the case of strawberry seeds, we studied possible interactions between PEMF, pre-freezing and pre-soaking of seeds immediately before planting. For bell peppers, extensive dose-response curves are reported. Total dose was calculated as D = R * t, where R = pulse rate (pulses per second) and t = treatment duration (seconds), the product yielding D (dose, i.e. total number of pulses in the PEMF treatment). Bell pepper seeds have an intermediate germination rate that allows the possibility for large and commercially important swings in germination rate both up and down, so we attempted to construct a full spectrum dose-response curve, beginning with PEMF dosing thought to be below threshold then extending into what was thought to potentially be the excessive and toxic range. We also attempted to identify which PEMF parameters are most relevant for dosing: total number of pulses versus “frequency” (pulse rate) and duration of exposure.

Results: Lettuce seeds showed evidence of a positive effect of PEMF on germination rate (79% unstimulated, 92% stimulated), with no effect on average time to germination. Strawberry seeds showed little effect of PEMF on germination, if any, and no apparent interaction with pre-freezing or pre-soaking of seeds. Bell Pepper seeds showed a clear inverse hormesis response within the dose range studied, and suggested a tri-phasic response to PEMF exposure for doses in excess of those applied. Specifically, extremely low PEMF doses had a negative effect on bell pepper seed germination rate, whereas intermediate doses had significant positive effects on germination rate, and germination rate began trending down again for the highest PEMF doses, suggesting a third phase of inhibited germination for extreme PEMF dosages. This latter trend was not fully elucidated by the present data set, but is only suggested by data points at the most extreme upper dosages included in this study. Overall, these results were entirely unexpected and may be of importance, therefore independent replication of these results is advised.

To explain the unexpected results, a detailed discussion of various multi-phasic dose-response curves is also included.

All negative findings were included in this report, specifically to counteract the dominant practice of publication bias in the life sciences.


Dennis R, Tommerdahl A, Dennis A. (2020). Inductively Coupled Electrical Stimulation - Part 4: Effect of PEMF on seed germination; evidence of triphasic inverse hormesis. Journal of Science and Medicine; 3(1):1-44.


The beneficial effects of PEMF on animal tissues have been reported for many decades [1,2]. For animal tissues, something on the order of 1000 peer-review papers have been published since the late 1950s, including a large number that remained inaccessible behind the Iron Curtain until the 1990s. It is also notable that PEMF is reported to have persistent beneficial biological effects on animal tissue which can last for months or years [3,4]. While a great deal of the scientific research even until recently has been built upon PEMF devices developed decades ago, more recent work has focused on determining the specific aspects of the PEMF pulse waveform and the identification of other PEMF parameters that have demonstrable and repeatable biological effects [5-8].

The large number of peer-reviewed publications on the topic of PEMF might suggest that we have accumulated a lot of useful and reliable information over the decades. But typical reports on the biological effects of PEMF are plagued by inadequate methodological detail, rendering the vast majority of such papers irreproducible, and thus, of questionable scientific value. Upon detailed review of more than 660 PEMF-related peer-reviewed papers, the author found that only 3% had sufficient methodological detail to be replicated (RG Dennis, unpublished). For example, it has been typical practice to report only the least relevant PEMF parameters, or to incorrectly report or omit key PEMF parameters such as waveform shape and rate of change of the magnetic field (dB/dt) [6-8]. Most alarmingly, while in some well-accepted fields of biomedical research it still remains possible, with some difficulty, to publish negative results, in generally marginalized fields such as PEMF, the publication of negative findings is virtually impossible [9]. Thus, the published body of literature, some 1000+ papers, should be interpreted as only the very tip of the iceberg, representing with high probability a larger number of research projects where PEMF was studied, but the results were negative. The magnitude of this is certainly in the range of 5,000 to 10,000 unpublished negative results on the use of PEMF, but given the generally chilly reception of PEMF in mainstream biomedical science, this number is undoubtedly much larger.

The problem with this of course is that it leaves us with woefully incomplete knowledge regarding the effects of PEMF on biological systems. We have, for example, no sense for the scope of PEMF types and settings that have been tried but simply did not work. It would be exactly this kind of knowledge that would be most helpful when searching the enormous parameter space of PEMF to find the range of parameters that confers biological benefits. It would be exactly analogous to searching a large house for a set of car keys, but having no way to keep track of places that have already been searched, resulting in a lack of knowledge of where the keys are not to be found. This will, of course, lead to a lot of re-searching of places that will turn up empty, and a great deal of frustration and lost time. Such a foolish search strategy, of blindly re-searching because we institutionally forget that which does not work is the acme of folly. Perhaps this is why we call the act of endlessly doing so research.

And yet, the finding of beneficial effects of some forms of PEMF on some types of living systems remains irritatingly persistent. In fact, it seems to pop up anywhere someone has bothered to look. The question emerges: are the effects of PEMF generally applicable to all living systems, or do they relate to only a narrow range of mechanisms in a narrow range of tissue systems, or only in a small subset of the grand diversity of life on earth? It would seem logical to hypothesize that if it could be shown that PEMF had beneficial effects over a wide range of life forms crossing the boundaries of not just species but of the kingdoms of life, then the basic mechanisms of PEMF within these living systems should be fundamental to life, and not just an unusual response of only the most advanced life forms. The most obvious question then is: does PEMF work on plants? If so, then the fundamental mechanisms which allow PEMF to benefit living organisms probably reach far back to a time before plants and animals took separate evolutionary paths. This would have been long before the development of eyes, or nervous systems, or flowers.

As it happens, there is good evidence that PEMF can benefit some of the fundamental mechanisms of plants. Reports of the effects of PEMF on plants are considerably less numerous than those on animals, totaling somewhat less than 100 when a thorough search of the literature was recently performed by the author (RG Dennis, unpublished). Many have focused on the study of the effects on seed germination, as recently reviewed by Pietruszewski and Martinez in 2015 [10]. It is entirely unclear whether the underlying biophysical mechanisms of PEMF in plants and animals are related. And the reports on the effects of PEMF on plants has all of the same deficiencies as the literature as it relates to animals. Nonetheless, the agricultural potential and commercial potential could be significant if PEMF confers benefits to plants related to seed germination, plant growth rate, harvestable tissue mass and or quality, resistance to environmental stressors and pests, reduction in the need for fertilizers or water, etc.

Related to plants, most reports focus on such a narrow range of PEMF interventions, or use germination methods that are neither practical nor scalable, that their results are of limited value to anything but basic research or the smallest scale agricultural endeavor. Therefore, our primary goal in this initial study was to identify PEMF parameters that would be of both scientific and general, large-scale use. To do this, first we attempted to determine if we could reproduce the general observation that PEMF has a significant beneficial effect on seed germination, and to record enough methodological detail so that our study could be extended in scope, replicated exactly, or replicated substantially but with variations (such as by the use of different seed types).

Our secondary goal was to explore the dose-response characteristics for PEMF exposure, in an attempt to give some insight into the PEMF activation kinetics to elucidate the fundamental PEMF parameters of interest, suggest potential biophysical mechanisms, and to identify where interactions may or may not occur between several commonly applied germination procedures. The over-arching approach of this study was to measure the effects of PEMF on a living system (seeds) using a clear and unambiguous biomarker (germination rate), with extremely well-defined PEMF parameters applied over a very wide range of doses. Such a study is generally not possible with animals or humans because of the large numbers of samples that are necessary, the extreme range of doses applied, and the lack of clear, easily attained, quantitative biomarkers for the effects of PEMF on animals. For this reason, our objective was to explore primarily this one biomarker (germination rate) of PEMF on living systems, and to begin to elucidate the dose-response relationship and begin the key PEMF parameter identification. The resulting insights were intended to inform future studies of PEMF on animals as well as plants, the search for optimal PEMF parameters and dosing, and studies directed to elucidate underlying biophysical mechanisms

The initial strategy of this study on seed germination was to replicate substantially (not exactly) and extend the general findings of a large body of scholarly work reported by diverse investigators, beginning with a report from the mid-19th century [11]. Generally, upon review of the literature [10], it is widely but not uniformly reported that certain types of electro-magnetic stimulation (alternating magnetic fields) applied to seeds prior to planting have a reliable and positive effect on seed germination rate for a wide range of plant species of agricultural importance, whereas other types of stimulation (static magnetic fields) typically have no repeatable, measurable effect, though this is not universally true. The exact PEMF parameters used, the key parameters necessary for the biological effect, direct comparisons of the effects of PEMF on seeds of differing germination rates, and dose-response relationships have not been well described.

In their 2015 review of the literature [10], Pietruszewski and Martinez present several equations to predict the effects of various magnetic field parameters such as “frequency” and “Gauss” on seed germination and growth. While these predictive equations are of course quantitative and appear to be quite accurate, they are not mechanistic, they are simply curve fits for reported data, and as such give no insight into whether or not specific magnetic waveform parameters that were neither controlled for nor independently varied were at the root of the biological effect. It should be noted that this is a weakness of the overall literature in this area of study and does not reflect negatively upon the review itself. Understanding this limitation of the literature which they review, Pietruszewski and Martinez point out that the precise mechanisms of the biophysical transduction of PEMF, while the subject of many hypotheses over many decades, remain largely unclear.

Therefore, in this study we carefully measure, control, and report the PEMF parameters (ICES®-PEMF). Based upon our earlier work [7,8] we propose that only one specific aspect of the magnetic waveform is necessary to elicit the observed biological effects, and we attempt to demonstrate that other aspects of the parameters (specifically frequency and dose) may be varied over a broad range without loss of biological effect. We hypothesize that the main contribution to the biological effects of magnetic field exposure of seeds prior to planting (and biological systems in general) arise primarily from the mechanism of electro-magnetic induction, and therefore that the DC or nearly DC components of magnetic exposure waveforms can be entirely removed from consideration. This suggests that the focus of study should be primarily on those components of the magnetic waveform that are changing in time above a certain threshold rate (dB/dt), that can be shown to be sufficient to induce an electrical current at or above that necessary to elicit a biological response.

ICES®-PEMF employs only those narrow portions of a sine wave where the rate of change of the magnetic field is sufficient to induce an electrical current above threshold and of sufficient duration to elicit the desired biological effect [7,8]. By eliminating the non-efficacious portions of the PEMF waveform, this approach eliminates as much as 99.8% of the total energy involved in each magnetic pulse when compared to pure sinusoidal waveforms [6,7]. The practical consequences are many, and the scientific ramifications include the fact that previous scientific work has been accounting for levels of energy that might have no direct bearing on the beneficial biological response of the system. Therefore, if we are able to replicate the general findings of the published literature (usually reporting improved germination rate) and are able to show this effect with a greatly reduced and tightly specified magnetic waveform [6,7], and we can show either a clear dose-response effect or an interaction with other processes that are thought to enhance seed germination, such as pre-soaking or pre-freezing for some types of seeds, then this suggests that this direction of study, using a very low energy, well-defined, non-sinusoidal magnetic pulse optimized for electro-magnetic induction, is worthy of further detailed exploration.

As this study reports initial findings to replicate an effect widely seen in many types of plants, without exactly duplicating any particular previously reported experiment, it is relevant to explain briefly why each type of seed was selected for this experiment.

Selection of Seeds

The seeds selected for this study were lettuce, strawberry, and bell pepper. These seeds represented a range of garden-variety edible plant types of commercial importance, with widely-ranging typical rates of germination: Lettuce at about 75 – 80%, bell pepper at 30 – 40%, and strawberry, which is typically well below 30%.

Certainly, there are more plant species of agricultural importance than there are academic papers on the effect of PEMF on plants, which number only in the dozens. There are several studies on crops of significance, such as wheat [12], tomato [13], onion [14], broccoli, coffee cultivars [14], and several others. Of direct relevance to the plant seeds studied in this report is the work by Izmailov et al. [15] on berry crops, specifically strawberry. But many of the published studies of PEMF on plants do not directly consider specific plants of commercial and agricultural importance. For example, many papers focus on a weed (Arabidopsis) of no real commercial or agricultural value, but rather one that is used as a scientific model because of its history and usefulness in genetic studies [16], while other recent papers are reviews of earlier work with a focus on finding “optimal” stimulus parameters [10], and others only indirectly relate to PEMF with a focus on developing molecular models of signaling and electrophysiology [17]. It is safe to say that the effects of PEMF on plant seeds and on mature plants and agriculturally important plant components for most commercially important crops remains largely unexplored. This may be because it has been assumed that the effects of PEMF would be more or less the same on all plants, and that findings from one could be generalized to all crops. Currently, there is no reason to believe that this is the case.

Lettuce Seeds

Lettuce seeds are known to germinate well and the plants themselves have considerable commercial and agricultural value. Lettuce was chosen because it was hypothesized that seeds with an already high germination rate would not show much improvement when exposed to PEMF, simply because there was not a lot of room for improvement of the germination rate. However, it remains possible that PEMF exposure may have deleterious effects on seed germination, and these effects might become more evident for seeds with normally high germination rates.

Strawberry Seeds

Formally speaking, strawberry “seeds” are not really seeds, but rather are the fruit of the strawberry plant:

“The tiny fruits actually contain the seeds. These seed-containing fruits are called ‘achenes’. An achene is occasionally also referred to as an ‘akene’, ‘achenocarp’, or ‘achenium.’” (

And the “strawberry” is itself not a true “berry”. Technically, it is an aggregate accessory fruit. ().

We anticipate that botanical taxonomy purists would have set this manuscript aside long ago without further consideration, on the basis that any use of the term “strawberry seed” disqualifies the work as nothing more than low-brow tripe. However, it is quite reasonable to consider the scientific possibility that strawberry “seeds” may react very differently to the exposure to a magnetic field stimulus than true seeds, or an already-germinated seedling, or a bean or legume, or even graftage, for example. In such cases, the potential target mechanisms (such as enzyme systems) may be complete and active, whereas in other cases (for true seeds in the dry condition, for example) this might not be the case. So, the selection of strawberry seeds was as much a result of decisions based on both low germination rates as on the fact that strawberry “seeds” are fundamentally different from true seeds. In the event that similar biological effects are observed, there would be the suggestion that the transduction mechanisms for seeds are fundamental to all living systems rather than just certain elaborated plant structures. If the biological effects are very different, it may suggest that PEMF acts on certain types of biophysical mechanisms while not acting upon others.

Prior work with strawberry seeds [15] had shown an improvement in germination rate of 14% over control, with an increase in seedling plant shoot weight of 33%, when PEMF was applied at 16 Hz for 6 minutes. But it should be noted that this paper used a different variety of strawberries (reported as “strawberry, garden”, not otherwise specified) than used in this study (wild strawberry, also known as Alpine Strawberry, Fragaria vesca), and their reported germination rates ranged from 34% to 48% (control vs. optimal stimulation), considerably higher than the 8% to 28% found in this study. Also, they were not clear about other pre-planting processes that were used on their strawberry seeds, such as pre-freezing or pre-soaking.

Jupiter Bell Pepper Seeds

The seeds from this plant have variable germination rates, and in our experience the germination rate is only about half that typically advertised by the seed provider. Bell peppers are a relatively valuable fruit, and an improvement in the actual germination rate may have commercial importance, as we feel this rate could be doubled and still be within practical limits. Bell pepper seeds in this study represent the intermediate range of germination rates, typically 30% to 40%, and therefore it was thought that the effect of PEMF on germination rate could be observed clearly whether the rate was increased or decreased by PEMF exposure, making this model suitable for dose-response experiments where PEMF exposure could be varied over a range that was thought to include both beneficial and potentially damaging levels of PEMF exposure.


Seeds were sourced from common Internet vendors and planted as follows:

Lettuce Seeds

  1. Ruby Leaf Lettuce
  2. Expected germination rate: > 75%
  3. Source: (Open Seed Vault)
  4. Seeds were soaked for 1 hour in reverse osmosis (RO) water at room temperature immediately before planting. Floating seeds were discarded before planting. Seeds were divided into two experimental groups based upon PEMF exposure (P+ or P-)
  5. Soil: Espoma organic potting mix, seeds gently pressed ¼” into soil and then covered

Strawberry SeedsAlpine Strawberry, fragaria vesca

  1. Alpine Strawberry, fragaria vesca
  2. Expected germination rate: < 30%
  3. Source: ()
  4. Seeds were treated in several groups prior to exposure to PEMF:
    1. Frozen at -4 ˚C for 27 days (or not: F-), allowed to thaw 1 day at room temp
    2. Thaw rate was slow, with the seeds remaining in the package.
    3. Pre-soaked in RO water for 1 hour (or not: S-) after freezing
    4. Seeds pre-treated with or without freezing (F+ or F-) and pre-soaking (S+ or S-) were then subjected to PEMF treatment (P+ or P-)
  5. Soil: Sifted Miracle Gro organic potting mix, seeds gently pressed ¼” into soil and then covered

Jupiter Bell Pepper Seeds

  1. Pepper, Jupiter, Sweet (Bell), 46134 B Organic, Lot# AK895
  2. Expected germination rate: stated = 89%, but only 30 – 40% based on our experience
  3. Source: Southern Exposure Seed Exchange
  4. Seeds were not pre-soaked prior to PEMF exposure or planting. In an initial experiment, seeds were separated into two experimental groups, one with no PEMF exposure (P-), the other with 30 minutes of PEMF exposure at 15 pulses per second (P+), for a total of 27,000 pulses. In a subsequent experiment, a range of PEMF exposures was used to study the dose-response characteristics of PEMF exposure as described in detail below.
  5. 'Soil: Sifted Miracle Gro organic potting mix (sifted through 3/16” hardware cloth), seeds gently pressed ¼” into soil and then covered.

PEMF Exposure

For each type of seed in this series of experiments, seeds in the PEMF+ groups were exposed to ICES® - PEMF using a Micro-Pulse model B5 pulse generator (), with a specially-built and calibrated solenoid coil array as shown in Figure 1 and Figure 2.

Figure 1.System used for generating ICES®-PEMF signals (Micro-Pulse model B5, Chapel Hill, NC)

Figure 2.Seeds were exposed to ICES® - PEMF by placing them into a polycarbonate sample cup, which was then placed into a solenoid coil assembly as shown. The assembly was held together with strips of non-metallized duct tape. Seeds were treated at a location inside the solenoid coils so that the magnetic fields were quasi-uniform for all seeds during treatment.

PEMF Pulse Generator Configuration and Calibration

The Micro-Pulse Model B5 programmable PEMF pulse generator was used for this entire series of experiments. A special set of coils was made by hand, without the use of the standard rubberized coating, so that the coils could be aligned and stacked axially to form a solenoid into which the seeds could be placed for the application of uniform PEMF pulses (see Figure 1 and Figure 2). The coils were wound exactly the same as the commercial production Micro-Pulse coils: 40 turns each of 28 AWG standard magnet wire, circularly wrapped to form individual coils of 1.50 inch (38 mm) internal diameter. Each of the four channels of the model B5 pulse generator was connected to drive two such coils in parallel, again identical to the coils used in the commercially available Micro-Pulse systems. In this way, the PEMF system used was electrically identical to the standard model B5 commercial product with four standard coil pairs.

The coils were aligned in parallel magnetically, coaxially, to form the final solenoid comprised of 8 coils.

The model B5 was then configured as follows: 15 pulses per second (pps), mode = hold (no changes or built-in delays) to generate a steady 15 pps bipolar alternating pulse train with electronic drive signals of 100 microseconds in duration. The pulse rate was verified on an oscilloscope to be 15.00 pulses per second. The B5 output intensity was then set to “13”, on the full scale of 0 to 15. The resulting individual magnetic pulse waveform and magnitude is shown in Figure 3.

The only change to this configuration that was made during this series of experiments was to change the pulse rate from 15 pps to 3 pps for comparison of the effect of pulse frequencies, as indicated in Table 1.

Lettuce seeds

Pre-soaked seeds were immediately exposed to PEMF (P+) at 15 pulses per second (pps) at a peak intensity of 255 Gauss for a duration of 10 minutes (Figure 2 and Figure 3). The control group (P-) was not exposed to PEMF prior to planting.

Strawberry seeds

Strawberry seeds were prepared in 3 groups, where each group was exposed to freezing (F+ or F-), soaking (S+ or S-) and PEMF (P+ or P-). In all cases with PEMF exposure (P+), the seeds were exposed to 15 pulses per second (pps) at a peak intensity of 255 Gauss for a duration of 30 minutes (Figure 2 and Figure 3). The control groups (P-) were not exposed to PEMF prior to planting.

Jupiter Bell Pepper seeds

Bell pepper seeds were exposed neither to freezing or soaking before planting. The experiment with Bell Peppers was done in two phases. In the first phase, there were only two experimental groups: PEMF exposure (P+), and no PEMF exposure (P-). Exposure was 15 pulses per second (pps) at a peak intensity of 255 Gauss for a duration of 30 minutes (Figure 2 and Figure 3). The control groups (P-) were not exposed to PEMF prior to planting. In subsequent dose-response experiments, PEMF exposures were added, to build the following set of PEMF exposure dosages. Note that the pulse rate drops from 15 pulses per second (pps) to 3 pps for the final four experimental groups italicized in Table 1. A wide range of exposure times and total dosage (numbers of pulses), some matching and some not matching between groups, was used because data of this type had not been previously published, and the nature and magnitude of the dose response was entirely unknown at the outset of this experiment.

Pulse rate (pps) Exposure Duration Total Pulses Peak Gauss
0 (P- controls) ZERO ZERO 0 (P- controls)
15 pulses per second 1 minute 900* 255 Gauss
15 pps 5 minutes 4,500** 255 Gauss
15 pps 30 minutes 27,000*** 255 Gauss
15 pps 1 hour 54,000 255 Gauss
15 pps 2 hours 108,000 255 Gauss
15 pps 5 hours 270,000 255 Gauss
15 pps 24 hours 1,296,000 255 Gauss
3 pps 1 minute 180 255 Gauss
3 pps 5 minutes 900* 255 Gauss
3 pps 25 minutes 4,500** 255 Gauss
3 pps 2.5 hours 27,000*** 255 Gauss
3 pps 72 hours 777,600 255 Gauss
Table 1.For each experimental group of 240 pepper seeds, the PEMF pulse rate (pulses per second, abbreviated ‘pps’), exposure duration (minutes and/or hours), total number of electromagnetic pulses, and peak Gauss delivered in each pulse is shown. Note that for several experimental groups, the exposure times were the same for 15 pps and 3 pps but the total number of pulses differed, whereas for other groups, the exposure time differed, but the total number of pulses was the same (matching pulse pairs are identified with * and ** and ***).

Figure 3.Each electrical pulse of 100 μs duration (applied from 0 μs to 100 μs in the figure) resulted in an approximately triangular magnetic pulse (green trace), which peaked at about 255 Gauss. The first time derivative of the magnetic pulse (dB/dt) is shown in blue (plotted using a rolling median filter of rank = 10), demonstrating that the value of dB/dt, the component of a magnetic pulse which causes electric fields by induction (Faraday equation), initially peaks at about 423 kG/s for the initial 7 μs of the pulse, then rolls off to an average of approximately 200 – 250 kG/s during the remainder of the 100 μs rising phase of the magnetic pulse. The falling phase of the magnetic pulse has a less pronounced initial peak, but a larger magnitude average slew rate of approximately – 300 kG/s over a period of approximately 80 μs as the current in the coil passively relaxes to zero. Pulse polarity was alternated, with each pulse being of opposite polarity to the one preceding it, and equally spaced at the stated pulse rate.

After PEMF exposure, seeds were placed ¼” deep into sifted potting soil in germination trays according to the seeding patterns shown in Figure 4.

Figure 4.Seeding density was Lettuce: 1 seed per cell, Strawberry: 4 seeds/cell, Peppers: 5 seeds/cell, with seeds carefully placed in each cell as shown. Seeds were placed by manual pick-and-place using a ¼” birch wood dowel, sharpened to approximately the diameter of each seed type, and moistened with a drop of water to allow individual seeds to be picked up and placed using surface tension, without damage to the seeds.

Germination was observed and recorded as follows:

Lettuce seeds were observed daily to determine the number of days to germination after planting (observed daily at the same time to assure 24-hour periods between observations).

Strawberries were observed and germinated cells were recorded at three pre-determined time points: at 2 weeks, 3 weeks, and 4 weeks after planting.

Peppers were observed and germinated cells were recorded at three pre-determined time points: at 1 week, 2 weeks, and 3 weeks after planting.

Growing Conditions

For germination and growth of the seedlings, temperature, light, and humidity were controlled, and germination domes were used (SoLight 48 cell domed germination trays). Room temperature and humidity were controlled tightly. The room was maintained at 73 ± 2 ˚F and 55 ± 5% relative humidity. All seeds were grown in 2 x 3 cell germination trays of vacuum formed polystyrene arranged into larger domed trays of 6 x 8 germination cells. Lettuce was grown 1 seed to a cell, strawberries 4 seeds to a cell, and bell peppers 5 seeds to a cell, as shown in Figure 4.

Each germination cell was misted every day with RO water throughout the experiment, and the domes were otherwise kept closed to retain maximum moisture throughout germination. Within the germination domes, temperature and humidity were monitored continuously. The temperature was measured to be in the range of 77 ˚F to 82 ˚F, and the humidity ranged from 84% to 91%.

Lighting Conditions

Natural light was excluded from the growing area. Seeds were illuminated by LED grow lights (MARS HYDRO SP150 LED Grow Lights 2x2 ft Coverage Full Spectrum Grow Light) on a 16-hour ON, 8-hour OFF light cycle. Lamps were positioned 18 inches above the germination domes, and the incident light intensity was measured to be in the range of 14,000 +/- 5000 Lux. Germination domes were shifted in their positions under the grow lamps daily to ensure that, on average, all germination cells received approximately the same amount of light over time. The spectrum for these grow lights is shown in Figure 5.

Figure 5.Spectrum for the grow lights used in this series of experiments. Product specifications provided by the manufacturer (Mars Hydro).



There were 72 seeds in each group (PEMF +, and PEMF –). As shown in Table 2, the germination rates were 92% for the stimulated (PEMF+), versus 79% for unstimulated (PEMF-) controls. The resulting z score is 2.125. The value of p is .03318, which is significant at p < .05. The mean number of days to germination and the standard deviation are given in Table 2 and Figure 6. A histogram showing the day when each seed germinated in each group is shown in Figure 7.

Germination Seeds 66 57
Germination Rate 92% 79%
Mean Germ. (days) 8.2 7.8
Std. Dev. (days) 3.0 2.7
Table 2.Lettuce seed germination rate with (PEMF +) and without (PEMF -) electro-magnetic stimulation.

Figure 6.Average (mean) days to germination of lettuce seeds, and the standard error for days to germination for stimulated (PEMF +) and unstimulated (PEMF -) lettuce seeds.

Figure 7.A histogram showing the day of germination for each lettuce seed. There is no apparent difference in the pattern or skew of the day upon which germination was observed between the two experimental groups.

At 20 days, each lettuce sprout was harvested and cut at ground level to measure the mass of plant tissue comprised of stem and leaves (not roots). Average wet mass was measured and is shown along with the standard deviation in Figure 8. The mass of the different sprouts was not distributed normally, and in any event, there were no differences in the wet mass of either statistical or practical significance. Dry mass would have been more instructive than wet mass, to allow a measure of total solid content, but was not possible due to the very small size and mass of most of the lettuce sprouts.

Figure 8.Wet mass (mg) of lettuce sprouts at 20 days after planting. No significant difference in means was observed.


The strawberry germination rate data is summarized in Table 3, with a more detailed analysis in the Appendix.

2 weeks 3 weeks 4 weeks
n % n % n %
Freeze+ 214 12% 308 17% 342 19%
Freeze- 290 16% 383 21% 394 22%
Soak+ 188 10% 279 15% 304 17%
Soak- 316 17% 412 23% 432 24%
PEMF+ 264 15% 346 19% 374 21%
PEMF- 240 13% 345 19% 362 20%
Table 3.The total set of 1,824 strawberry seeds were pre-treated with freeze (+ or -), soak (+ or -), and PEMF (+ or -), then planted and germination observed at 2, 3, and 4 weeks to determine the germination rate at each time point. Note that each seed was pre-treated with a combination of F, S, and P (+ or - each), for 8 groups: F+ S+ P+, F+ S+ P-, F+ S- P+, F+ S- P-, F- S+ P+, F- S+ P-, F- S- P+, F- S- P- , and these groups are aggregated in this summary table to show, for example, all seeds that were PEMF + (whether F + or -, or S + or -), and so forth.

From both a practical and statistical standpoint, there were no clear beneficial trends or differences of any agricultural importance between any of the experimental groups, but when looked at in isolation several minor interesting observations were made:

  1. Pre-freezing did not enhance germination rate, and may have suppressed it slightly. This is contrary to conventional wisdom when planting strawberry seeds, and pre-freezing is often stated as a requirement prior to planting. Yet our finding was that pre-freezing appeared to reduce the germination rate slightly but not significantly at 2, 3, and 4 weeks (all time points).
  2. Pre-soaking also did not have a positive effect on germination rate. Our data suggested that pre-soaking also resulted a slight reduction of the germination rate of strawberry seeds at 2, 3, and 4 weeks.
  3. All else being equal, with the data viewed in aggregate for all groups, PEMF + may have resulted in a slight improvement in germination rate compared to PEMF -, if only at week 2 (but not weeks 3 or 4), and only by a very small and statistically unimpressive amount. This difference may (or may not) be real, but nonetheless is of questionable practical importance.
  4. Combinations of two or three of these interventions did not appear to have synergistic or significant effects when combined.

Details of the data and the statistical analysis of the data on strawberry seed germination are included as Appendix I.

Bell Peppers

For PEMF+ groups, bell pepper seeds were subjected to two different PEMF pulse rates (3 pps and 15 pps) and for exposure times ranging from 1 minute to 72 hours, but always at the same PEMF pulse intensity (Figure 3). Germination rate (%) was recorded at times of 1, 2, and 3 weeks. By 3 weeks, in many cases differences in germination rate were clearly evident by visual examination, as shown in Figure 9. The dose is plotted as the time duration of exposure (minutes, Figure 10 above), or alternately as total number of pulses (Figure 10 below). By so doing, for example, an exposure to PEMF at 3 pps for 25 minutes would be the same dose, in terms of total PEMF pulses, as an exposure of 15 pps for 5 minutes, because each exposure subjects the seeds to a total of 4,500 identical electromagnetic pulses, with pulses as illustrated in Figure 3.

Figure 9.Visual comparison of germinated bell pepper seeds at three weeks. Controls (left) with no PEMF exposure are visually contrasted with a PEMF exposure of 3 pps for 1 minute (center) which shows a decrease in germination rate, whereas 3 pps for 2.5 hours (right), showing increased germination rate, and what appears to be healthier, more uniform seedlings.

Figure 10.Germination rate of bell pepper seeds at three time points after planting: 1, 2, and 3 weeks (color coded), and for each color the lighter shade represents 3pps, whereas the darker shade, represents 15pps. The dose-response is shown for exposure times ranging from 1 minute to 72 hours (above), resulting in total electro-magnetic pulse exposures ranging from 180 pulses (1 minute @ 3 pps) to 1,296,000 pulses (24 hours at 15 pps) (below). Note that while the dose-response data is plotted on a log scale for exposure (time or number of pulses), these are plotted against the germination rate for control (PEMF -), expressed as zero exposure time or zero pulses, shown both by a zero value on a broken-log scale near the origin, as well as a dashed line indicating the germination rate at each time for seeds without PEMF exposure (PEMF-). Each of the three baseline germination rates is color coded to correspond to the relevant time point (1, 2, or 3 weeks).

Considering the dose-response data points that are most relevant for agricultural purposes, as a preliminary observation we consider whether any particular dosage is likely to result in a significant change in the germination rate, positive or negative, when compared to no application of PEMF. The most commercially-relevant data may be summarized as shown in Table 4.

900 Pulses 27,000 Pulses
Control (PEMF-) 3 pps 15 pps 3 pps 15 pps
TOTAL SEEDS 240 240 240 240 240
Germinated Seeds 168 95 112 208 218
Germination Rate 70.00% 39.60% 46.70% 86.70% 90.80%
Z Value (2-tailed) - 6.6948 5.1846 -4.4317 -5.7509
p < … - 0.00001 0.00001 0.00001 0.00001
Table 4.Germination rates for Jupiter bell peppers at the three-week time point, showing the values for control (no PEMF, or PEMF-), the greatest level of germination suppression (at 900 total pulses) and greatest enhancement of germination (at 27,000 pulses) for which both data sets (3 pps and 15 pps) have data. Exact maximum and minimum pulse numbers would have differed from these. Z values test only the hypothesis that each dose condition is not different from control (no application of PEMF). This does not test differences between dosages.

From the full tabulated data set, the following set of maximum germination rates emerge (Table 5). Note, these are not “optimal values” because it is not possible to test all potential parameter values. Thus, optimal values invariably fall between the measured points near the maximum. Also note that these findings should be viewed as preliminary. Based on the available data, the greatest increase in germination rate was achieved by applying 15 pps for 75 minutes (67,500 total pulses) for week 1 and 2. At week 3, the maximum germination rate occurred for 15 pps at 30 minutes (27,000 total pulses), but this value was only slightly above the germination rate at 15 pps for 75 minutes at 3 weeks, which was 29% more than baseline (not shown) rather than 30% as shown in the table for 30 minutes of stimulation at 15 pps. As shown in Figure 10 though, the maximum germination rates occurred over fairly broad peaks, so it is unlikely that there is one optimal value, rather a range of values would best describe an “optimal range”, with only small changes in the germination rate around the maximum for at least ± 50% of the stimulation time. Thus, the maximal germination rate probably occurs for stimulation times somewhere between 30 minutes and 2 hours, for 15 pps stimulation pulse rate.

Maximum Germ Rate Improvement over Baseline pps Stim Time
Week 1 51% 72% 3 2.5 hours
Week 1 52% 76% 15 75 minutes
Week 2 78% 39% 3 2.5 hours
Week 2 82% 46% 15 75 minutes
Week 3 87% 24% 3 2.5 hours
Week 3 91% 39% 15 75 minutes
Table 5.Maximal germination rates for each time point and pulse rate (pps) measured during this study. “Improvement over baseline” is the percent increase in germination rate over the baseline germination rate. Baseline germination rate increased at each time point, from 30% at week 1, to 56% at week 2, and up to 70% at week 3.

In summary, the application of PEMF at the dosages which appear to have the greatest effect on germination rate each cause significant, and potentially commercially important, changes from the baseline germination rate of 70% at 3 weeks. When only 900 electromagnetic pulses are applied, the germination rate drops from 70% (baseline) to ~40% (3 pps) and ~47% (15 pps), whereas higher dosages of 27,000 pulses result in increased germination rates of ~87% (3 pps) and ~91% (15 pps). When comparing each dose with baseline, these changes in germination rate are highly statistically significant, and are probably of commercial/agricultural significance.


Overall, it did appear that the use of this specific form of pulsed electro-magnetic fields (ICES® - PEMF) had a clear effect on seed germination under some circumstances. Since ICES® - PEMF is optimized for inductive energy transfer, and has virtually zero steady magnetic component to the signal pulse wave form, these results suggest that the mechanism of action is one involving electro-magnetic induction (the component of a magnetic field that changes with time) rather than the steady component of a magnetic field, which is in general agreement with the body of literature on this topic stretching back at least a century.


The high germination-rate lettuce seeds would normally have been expected to germinate at a rate of approximately 80%. The unstimulated lettuce seeds germinated at 79%, which was to be expected, whereas the stimulated seeds germinated at 92%. A simple statistical analysis (z-score) suggests that this difference is significant, at least at the p < 0.05 level, but barely so. Had the numbers been off by a single germinating seed (one more non-stimulated germination, or one less stimulated germination), the result, while suggestive, would have failed the test of basic statistical significance. This appears to be typical of biological effects of PEMF: differences, if they do exist, are often not statistically impressive.

On the other hand, given that lettuce would normally germinate at a rate of about 80%, there is not much room for improvement; the germination rate could not be doubled, for example. And it is question for individual growers whether an improvement from 80% to 90% would have any agricultural or economic value, given the added cost and effort that would be associated with PEMF stimulation of seed stock on their agricultural scale. In the case of lettuce, the value of the effects of PEMF stimulation on seed germination is certainly of scientific enquiry, as well as in cases where every seed is valuable, such as in the case of colonies on distant moons or planets. It is also worthy of note that the PEMF applied to lettuce seeds did not have a negative effect on the germination rate, which was certainly a possibility. In light of the dose-response data reported below for bell peppers, it still may very well be the case that incorrect dosing may reduce, or have no effect on, lettuce seed germination.

With regard to the time of germination and wet mass of the germinated lettuce seedlings, there were no apparent differences between the two experimental groups (PEMF+ versus PEMF-). It is entirely possible that significant effects of PEMF on biomass may be more evident in the root system. But once again, this may depend upon PEMF dosing.


Germination rates in the current study from stratification (+ or -), pre-soaking (+ or -), and PEMF (+ or -) prior to planting were surprising. Our results suggest that little to no effect was seen with any of the three treatments, with a small decrease in germination rates due to soaking the seeds for 1 hour in RO water prior to planting.

Looking to how our results fit in with literature, we see that the task of increasing germination rate and/or decreasing time to germination has been reported in the literature going back to at least 1950 [18]. Due to the notorious difficulty of starting strawberry seeds, it is important to note the economic significance of the topic: although vegetative propagation is a much easier and effective means of producing strawberry plants, breeding new cultivars requires cross pollination and thus germinating seeds. Keeping that in mind, it can be seen why it is so important to improve germination. Having seedlings germinate at the same time allows for direct comparison in post-germination growth and development to help choose economically viable cultivars, as well as helping to prevent overlooking a phenotype that may germinate more slowly but has other attributes that are suitable for cultivation.

Both soaking and freezing prior to planting have been extensively reported in previous studies [18-25] with varying results. Interestingly, it seems that different cultivars respond differently to treatment protocols [21,22,24]. For example, when testing 5 different cultivars, Wilson et al. [24] found that chilling seeds for 4 weeks at 2oC prior to planting had no effect on 3 of the varieties, while the same treatment caused a significant improvement in germination rate of one variety and a significant decrease in total germination of the fifth variety. Furthermore, increasing cold exposure to 8 weeks significantly increased the germination rate in one of the varieties that did not seem to respond to a 4-week cold treatment. Thompson [23] concluded that a 6–8-week cold exposure should be used as a standard protocol in order to increase germination rates. Applying this to our data, we hypothesize that perhaps 1) our freeze time was not long enough to stimulate germination, 2) our cold exposure was too cold, or 3) the variety used in the present study is not affected by cold. It is also possible that germination rate would have been higher had we continued observations longer than 4 weeks.

The literature also provides some clarification why we saw no influence of soaking on our germination rates. Guttridge & Bright [25] investigated combinations of soak times, temperatures, as well as cold stratification and did overall observe a positive effect of soaking, with significant improvement germination speed from a 6-week soak compared to a 3-week soak. After following seeds out to 60 days, it was noted that the final germination rate was similar for all treatments, but soaking sped up the time to germination after planting. We note that perhaps our soak time was simply not long enough to cause any effect.

Overall, the general findings from strawberry germination papers are that there are huge variations in response to any treatment based on cultivar, and responses are significantly affected by the length/strength of the given treatment. Given that Izmailov et al. [15] found significant improvement in strawberry seed germination following PEMF exposure, it is quite possible that a different PEMF protocol in the present study could have yielded significantly different results and/or a different cultivar may respond differently to PEMF exposure.

The Importance of Negative Results

The life sciences face serious challenges from the replication crisis and more specifically publication bias. The negative results in this study, insofar as strawberry seeds are concerned, are but one small example of this much larger problem that is particularly endemic to the life sciences [26]. This is an enormous problem in the sciences because when there exists a strong bias to publish only positive scientific findings, their repetition eventually elevates them to be accepted as ‘fact’ [26,27]. The ratio of positive results in the peer-reviewed literature has recently risen to about 90%, and is probably higher, but when studied closely the number should be much closer to 50%, and for many areas of high-risk and innovative research, it should be much less than this [26-29]. This is all discussed in great detail in a Letter to the Editor of JoSaM [9], to be concurrently published with this report.

In a trend similar to the medical literature, it is our impression that negative results in the agricultural literature were fairly common in decades past, but the publication of negative results in the agricultural literature appears to be less common in recent years. But this phenomenon does not appear to have been as well studied in the agricultural literature as it has in the medical literature of the past two decades.

With all of this in mind, what does the set of data presented allow us to conclude? Our initial aim was simply to replicate and verify two simple assertions: pre-freezing for 3 to 4 weeks prior to planting is essential for strawberry seed germination, and pre-soaking immediately before planting is also helpful in this regard. We took these statements as “given” and endeavored to show whether these effects were additive and whether or not they interacted with, or were synergistic with, or perhaps could be further enhanced or replaced by, a brief exposure to PEMF. In the course of the experiment, we simply endeavored to replicate the basic prior findings and beliefs as a starting point.

While our findings are certainly not conclusive, we consider them to be highly suggestive that some of the commonly held “facts” about how one should plant strawberry seeds should be revisited with a skeptical eye. We would go further to say that this may indeed be the case for much of the received wisdom as it relates to gardening and agriculture in general. For any process that is widely accepted, the practice of which has significant impact on cost, complexity, effort, or outcome of any agricultural activity, we offer the Russian proverb Doveryai, no proveryai (Russian: Доверяй, но проверяй); “Trust, but verify”. If it is important, it may be worth additional or renewed scientific scrutiny. If it is supported it should be published, but especially if it can be shown to be false, by all means, publish the negative result.

And with all of this, it may still be true that PEMF, correctly applied at the optimal dose, may in fact improve strawberry seed germination. Probably not for the type and dosing of PEMF that we applied to strawberry seeds, but at least now we know that. And pre-freezing and pre-soaking do not seem to be helpful for getting improved germination for seeds from the Alpine Strawberry, fragaria vesca. For those who disagree, there is enough methodological detail in this manuscript, and detailed statistical analysis of the data in the Appendix, to formulate a counter argument.

Bell Pepper

In the simplest interpretation, it could be said that our initial experiment (P+ = 15 pps for 27,000 pulses, P- = no stimulation control) yielded very significant and commercially important results if independently verifiable, and that PEMF “works” to enhance Jupiter bell pepper germination rate. Many would incorrectly interpret this to mean that PEMF universally enhances bell pepper seed germination, and that this paper “proves” that it works, and that 15 “Hz” is the optimal frequency. Of course none of this is true, strictly speaking, but this is fairly typical of how PEMF results are interpreted, even by scientists in the field. It would be much more accurate to say that PEMF at one range of dosages, using the waveform parameters described above and a frequency chosen essentially at random as an educated guess, does seem to enhance pepper seed germination, while at lower dosages it may suppress germination, but it is unclear from this initial single-dosage experiment the exact range for the optimal dose, and it is also unclear which PEMF parameters must be closely controlled: frequency? exposure time? other? So, the initial single-dose experiment by itself is not very revealing, other than to further support the notion that, somehow, PEMF probably “works”. Subsequent dose-response characterization reveals much more, but still does not identify the “optimal” parameters, or irrefutable “proof” of anything.

With this data we do have the advantage of a much more replete dose-response curve, which sheds a great deal of light on how PEMF could potentially be harmful, beneficial, or have no detectable effect at all, depending upon the dosage. And we have gained insight into which parameters are probably most directly relevant to PEMF dosage. Further, the best range of dosages to search to determine the optimal beneficial dose becomes abundantly clear upon even a cursory examination of Figure 10, and the sensitivity of the optimal dose to changes also becomes clear: between 27,000 and 108,000 pulses there is probably a maximum, with little change over this fairly broad range of dosages. So it is safe to say that precise dosing is not essential to reliably achieve on the order of at least 90% to 95% of the potential benefit of PEMF for bell pepper seed germination.

The dose-response curves strongly suggest that PEMF can have both a beneficial and harmful effect on germination rate. Broadly speaking, the dose-response curves appear to indicate that the effect of PEMF exposure on germination rate is best described by using the total number of electromagnetic pulses (Figure 10, below), rather than the total exposure time (Figure 10, above). This is because of the basic shape of the curves, which are biphasic and initially indicate a suppression of germination rate for very low doses of PEMF, followed by increased germination rate for higher doses, with each phase of the curve achieving a minimum or maximum within the dosing span studied. The greatest suppression of germination rates occurred at approximately 900 pulses for both pulse rates (3 pps and 15 pps), and at all three time points (1, 2, and 3 weeks). Germination rates were suppressed by 40% to 47% when 900 pulses were applied to seeds before planting. The amount of suppression was consistently greater when 3 pps were applied, generally resulting in ~30% more suppression than 15 pps at 900 total pulses, compared to baseline.

The dose is expressed as the total number of pulses (D), which of course can be directly calculated by multiplying the pulse rate (R, in units of pulses per second) multiplied by the time (t) in seconds:

D = R * t

As a practical matter, it is easier to set a PEMF dosage by timing the exposure at a fixed pulse rate. So for the remainder of this discussion, we will consider setting the exposure time (t) with a fixed pulse rate (R) to get the desired PEMF dose (D), with the understanding that it is the total number of pulses being applied, not the exposure time or pulse rate per se, that defines the PEMF dosage.

The effect on germination rate at higher dosages of PEMF was positive, with an increase from 87% to 91% being typical at 3 weeks. This maximum occurred at a PEMF exposure of 27,000 total pulses, with the real maximum probably occurring at a total pulse number somewhat higher than that. The peak number of pulses for the 3 pps exposures appears to have been missed by the exposures that were included in this experiment, but probably occurs between the last two (two highest) exposures, somewhere between 27,000 and 777,600 pulses at 3 pps. Interpolating between these two points, it appears that the enhanced germination rate was approximately the same for 3 pps and 15 pps, and the peak exposure for 3 pps may have occurred with slightly higher numbers of total pulses than for 15 pps, but that is not absolutely clear from the available data. With this limitation in mind, the increased benefit at 27,000 pulses for 15 pps (90.8% germination rate) was only about a 7.5% improvement over the increase in germination rate for 27,000 pulses at 3 pps (86.7%), when compared to baseline (70% germination rate). With a more complete dataset, the difference in germination enhancement, and the optimal pulse rate and number of pulses at which this optimal difference occurs would be expected to be somewhat different than indicated by the individual data points we have presented.

Effect of Frequency

In summary and broadly speaking, the harmful effect of low dosages of PEMF exposure were somewhat greater at 3 pps than at 15 pps, whereas the beneficial effects at the comparable near-optimal dosage of 27,000 pulses was potentially greater for 15 pps when compared to 3 pps, but probably not by as much. It might be said therefore, that overall, PEMF stimulation at 15 pps was somewhat more beneficial than stimulation at 3 pps, at all dosages in the range tested except for the very highest dosages tested (Figure 10, below). This does however demonstrate convincingly that precise “frequencies”, or harmonics of Earth ionospheric resonances such as Schumann frequencies (~7.83 Hz primary), are entirely unnecessary and irrelevant to the mechanism of PEMF on plant seeds. We hypothesize that this finding is generally true for the effects of PEMF on all living systems. For the system studied, the PEMF dosage appears to be almost entirely a function of total number of pulses, and not one of “pulse frequency”, for any given PEMF waveform shape and magnitude.

“Optimal” stimulation values

The only pulse rates tested in this study were 3 and 15 pps, so it is difficult to make broad generalizations other than to suggest that higher frequencies in this approximate range appear to result in very slightly better germination rates, but this difference would need to be verified and may not be evident upon subsequent experimental replication. With the data from this study, at 15 pps stimulation, it appears that the maximum germination rate occurs somewhere in the range of 30 minutes to 2 hours. With only two or three data points falling in this range, it is difficult to narrow this time more precisely. But the peak is fairly broad and so it would appear to be safe to say that 1 hour of stimulation at 15 pps would reliably fall near the middle of the optimal range. Of course, this is shorthand and the real parameter of consequence is total number of pulses at a specified pulse rate. The total time of PEMF application (t) is reported as a summary finding simply because it is an easier value to apply in a practical sense and can be calculated directly from the pulse rate (R) and desired number of pulses (D) as indicated above.

Shape of the Dose-Response Curves

Though replete with suggestive information, these dose-response curves for Bell Pepper seed germination are best viewed as preliminary, as an attempt to identify trends that may emerge over fairly large ranges of dosing, and different methods of exposure of the seeds to equivalent doses, specifically: exposure time at different pulse rates (pulses per second, or pps), versus exposures as measured by different numbers of total PEMF pulses applied at different pulse rates, and therefore differing exposure times. Essentially, our goal was to begin to elucidate the general characteristics of the dose-response relationship for seed germination as a function of PEMF dosage. Bell pepper seeds were specifically selected for the dose-response experiment because of their intermediate germination rate, which allowed the detection of both positive and negative effects on seed germination with maximum range and resolution.

Dose-response characteristics

We initially expected that the dose-response characteristics might involve a minimum threshold dose, followed by a classic sigmoidal dose response. We primarily were thinking to set dosages to identify the threshold as well as the doses that would show the approach to the asymptote (peak response) of the sigmoid. The bi-modal dose-response curves that showed up in all datasets for pepper seeds, best illustrated by Figure 10, were entirely unexpected. Therefore, a discussion of dose-response curves is necessary to fully interpret our results.

The simplest and least realistic model of dose-response is the linear no-threshold model as shown in Figure 11. The a priori assumption of such models is that presumptively beneficial agents or environmental conditions are always beneficial, at all doses, and the benefit is directly and linearly related to the dose. Correspondingly, this model assumes that harmful substances or conditions are always harmful, at all doses, and that the level of harm is directly proportional to the dose. So far as we have been able to determine, there exists no such dose-response relationship in any living system, and it is safe to say that if such a dose-response relationship does exist in reality, it is certainly the exception and not the rule. Linear no-threshold models are most often employed when sufficient reliable data simply are not available, or in support of extreme views or overt misinformation campaigns such as those claiming that “no amount of [substance X] is safe.” Real dose-response relationships typically have threshold values below which no effect is detectable, and they tend to be non-linear, with a pronounced sigmoidal shape (Figure 12) and often with maxima or minima within their measured range (Figure 13, red trace).

Figure 11.Two linear no-threshold dose response curves are shown schematically (not real data). In such a dose-response model, when a benefit is seen (green) it begins immediately even at the lowest doses and increases linearly with dose. In the case of harmful agents (blue), the harmful effects are detected beginning with the lowest doses and the amount of harm increases linearly with increasing dose.

Figure 12.Typical sigmoidal dose-response curves are shown schematically (not real data), displaying the classic sigmoidal shape with implied thresholds at the measurement resolution limit where the curves approximate zero, along with thresholds of clinical or scientific significance, often set at 10%. Source: ToxTutor – NIH [30].

While sigmoidal dose-response curves are far more realistic than linear no-threshold models, an even better and more realistic model for many types of biological responses to stimulus is that of hormesis. When a chemical agent or environmental factor is beneficial at lower doses, but becomes harmful or toxic at higher doses, the dose-response is often characterized as ‘hormesis’, and the phenomenon is known as a ‘hormetic effect’. In the fields of biology and medicine hormesis is defined as an adaptive response of cells and organisms to a moderate (and often intermittent) stress. There are many common examples include ischemic preconditioning, exercise, dietary restriction, sun exposure, and exposures to low doses of certain phytochemicals [31].

A concise working definition of hormesis is: ‘a process in which exposure to a low dose of a chemical agent or environmental factor that is damaging at higher doses induces an adaptive beneficial effect on the cell or organism’ [31], or simply stated: a little bit is good for you, but too much is not. Hormetic dose-response curves are characterized by being biphasic: having distinct phases where either a benefit is observed, or harm is indicated. Hormetic dose-responses are remarkably prevalent in the environmental toxin literature [32], and Mattson goes on to state “thousands of published articles include data showing biphasic responses of cells or organisms to chemicals or changing environmental conditions... While not all toxic factors may induce a biphasic dose response in cells and organisms, many clearly do. Examples include many chemicals, temperature, radiation, exercise, energy intake and others.” [31].

An organismal way to view hormesis is eustress, meaning “good stress” which is a positive, beneficial form of stress. Too much stress can be bad, but a reasonable amount of the right type of stress can be very beneficial. The concept was first put into writing by the Greek poet Hesiod in the earliest known Greek literature almost 3,000 years ago, and this has become one of the cornerstones of Western Civilization: healthy competition (good stress) brings out the best in individuals, the open market, and civilization in general. Competition is essential for the evolution of organisms and Ideas. But too much stress causes injury and death. And hormesis is the biophysical analog of the concept of eustress.

Hormesis may also be cross-modal, where one type of hermetic agent can confer broad protection against several other types of stress, examples being the broad protective effects of physical exercise, diet restriction, or dietary phytochemicals. In each case, particularly with phytochemicals, the protective effect will often dimmish or disappear when studied at high doses [31].

The low-dose-benefit and high-dose-harm of a classic hermetic dose-response curve (Figure 14) does make sense when viewed from the perspective of the mechanism of adaptive low-dose stress to otherwise toxic substances. However, the inverse hormesis effect, as shown schematically in Figure 15, in which low doses are harmful or inhibitory whereas higher doses are beneficial, is less easy to explain mechanistically. Yet, that is precisely the effect that was evident in the dose-response of the bell pepper seeds to PEMF exposure (Figure 10).

The mechanisms and biological advantages of an inverse hormesis dose-response are less clear than those of simple hormesis. Nonetheless, observations of an inverse hormesis effect have been reported elsewhere [33].

For the bell pepper seed germination data, this resulted in a series of dose-response curves that, taken together, are not readily analyzed for statistical significance. However, clear and repeatable trends do emerge, and statistical analysis notwithstanding, these trends suggest several much simpler and more direct experiments that should be conducted to establish clear and repeatable effects, while also suggesting equivalency between certain PEMF stimulus parameters (total number of pulses rather than pulse rate or frequency).

The first trend that is clear in all dose-response plots is the fact that lower dosages (~900 total pulses) tended to reduce the germination rate, whereas longer exposure times (27,000 to ~180,000 total electromagnetic pulses) resulted in very clear increases of germination rate above control (PEMF-). It also appears that the effects on germination appeared to be broadly equivalent whether the PEMF stimulation was applied at 3 pps or 15 pps. The units “pps” (pulses per second) can be taken to be equivalent to the widely used term ‘frequency’ (Hz), as both describe the number of electromagnetic pulses that are applied each second.

Looking at the consistently evident trends, it is possible to make some general observations. From the available data, it would appear that PEMF exposures below ~10,000 total pulses may reduce germination rate, with peak suppression of germination occurring at ~900 pulses. But germination rate may be enhanced by as much as 25% to 40% when about 27,000 to ~180,000 pulses are applied immediately before planting the seeds. Even greater exposure may then begin to reduce the enhancement of germination rate. The range of dosages in this series of experiments did not extend far enough to indicate whether a second range of suppression would result in a clear over-dosing, which might have either brought the beneficial effect to zero (baseline), or may have reduced it further into a negative effect (reduced germination rate), thus creating a third phase, or tri-phasic dose-response curve.

Figure 13.Based on a graph originally presented in: Beneficial dose is determined by superimposing the effective dose and toxic dose curves. In this case, the resulting beneficial dose (red) represents the percent of test subjects or specimens that benefit at any given dose (% Effective – % Toxic).

Figure 14.A classic illustration of hormesis is shown schematically (not real data), where low doses are shown to confer some benefit, whereas increasing dosages tend to become harmful, toxic, or even possibly lethal. This would be typical of radiation hormesis, for example. Typically, the actual benefit may be much smaller than shown, but this represents the overall trend of the hormetic effect.

Figure 15.This diagram shows what an inverse-hormesis effect would be expected to look like, where the dose-response would be very similar to the inverse of a typical hormesis response, until such high doses were applied that an overdose condition begins to dominate, and the benefits of higher dosages begin to return to zero. Presumably, at extremely high doses (greater than those shown) the effects could again become harmful. It is important to note that in all hormesis-like responses, the magnitude of the beneficial response may be very different from the magnitude of the harmful response. Depending on the specific circumstances, the harm or the potential benefit may be much smaller (or larger) than the converse, when looked at over the full range of dosages.

Several things are important to point out as they relate to dose-response curves and hormesis. First, the vast majority of scientific studies do not report dose-response effects over the full realistic range of dosages. In fact, most report only one dosage compared to a control group (zero dosage), and then attempt to show that a substance “works”, i.e., is beneficial, or harmful, using a simplistic “A versus B” comparison of means. While the scientific literature is dominated by such over-simplifications, upon consideration of realistic dose-response curves, it is clear that the benefit or harm of any given substance or form of energy will depend as much upon the dose as upon the nature of the substance or energy being applied, thus justifying the adage: the dose makes the poison [34]. Second, the general shape of the hormesis response makes sense in light of the proposed mechanisms for such a response, namely, where modest or low stresses applied to homeostatic or enzyme systems may confer some benefit as they may elicit an adaptive stress response [31].

Interpretation of the Data

At the very least, the bell pepper seed germination data suggest that the selection of PEMF dose is important, but need not be precise, and may result in either reduction or increase of germination rate by as much as 30% above baseline, depending on a number of other factors. From a broader perspective, this may explain much of the confusion in extant data on the effects of PEMF in living systems, where a significant benefit (or no benefit) may simply be a result of the dosage being studied. Since the vast majority of PEMF studies do not look closely at dose-response, most selecting a single dose without justification, we might expect that many important biological effects of PEMF may have remained undetected simply because the dosage selected for the study was wrong; it was not in the beneficial range. This may be the case, for example, with the strawberry seed germination rates reported in this manuscript: failure to see an important positive effect may have been simply a matter of the initial selection of dose, since all strawberry seeds in our experiments were subjected to the same PEMF dose (or zero dose for control).

It is also important to note that the effect was largely independent of the frequency of pulsing. This suggests that the waveform shape and total number of pulses are more important than “frequency”. Waveform shape and total number of pulses relate directly to the total inductive energy that is applied to the seeds before planting. Frequency may only be a factor so long as it falls within a broad range. In this study, we only looked at 3 pps and 15 pps, but future studies could include a wider range of frequencies to verify that frequency per se is not a critical parameter in PEMF dosing.

It also suggests that phenomena related to resonance, which are highly frequency specific, are probably not fundamental, and may be entirely irrelevant, as they relate to the biophysical mechanisms of PEMF.

The data presented have numerous limitations, and these should be considered as well. First, these only apply to the specific seeds and experimental conditions described herein, and these dose-response relationships may not reflect the general response of seeds of all types. The optimal dose for enhanced germination rate should be established for each seed type and germination condition being considered, until large amounts of data have been collected, and general trends emerge. Dose-response curves can be suggestive, but statistical analysis is complicated by the fact that each data point plotted is simply the germination rate for one relatively small group of seeds, so error bars are difficult to establish at each data point.

Perhaps most important, it is essential to replicate these general results independently before any categorical statements should be made about the presence of the putative “inverse hormesis” effect.

The presence of an inverse hormesis effect was certainly unexpected and may indicate a fundamental difference between seeds and adult plants or animals to PEMF exposure. That is, this may point to a fundamental mechanism that pertains more to seeds than to organisms at other stages of development.

Looking more carefully at Figure 10, one additional trend is suggested. At the risk of going beyond the data at the highest dosages, the germination rates clearly begin to again fall toward or below baseline. The data do not extend well into this high-dose region, but the trends in all sets of dose-response data suggest a third phase in which excessive PEMF dose levels begin once again to have less of a positive effect, where germination may drop to very low numbers, or even may again become suppressed, due to excess PEMF exposure. The general effect may therefore be tri-phasic. This could be described as the “Goldilocks Effect” in dose-response:

Too little = BAD

In the middle, just the right dose = GOOD

But too much = BAD

Strictly following the definition of “hormesis”, tri-phasic hormesis responses would begin with beneficial effects at very low dosages, with harm being shown at intermediate dosages, and then benefit at higher dosages. These might be described as “good-bad-good” dose-response curves. Tri-phasic hormesis responses (good → bad → good) have been reported for zebrafish for adaptive responses to radiological exposures [35,36] and for cancer growth [33]. The inverse would be a triphasic response that shows just the opposite effect with increasing dosage: harm → benefit → harm, or “bad-good-bad”, where the beneficial dose is in the middle, and too much or too little causes harm. This is illustrated in Figure 16.

Figure 16.A proposed tri-phasic inverse hormesis effect. Note that the effect would generally have three phases, but that one or more of the harm or benefit phases might be very small compared to the scale shown. For example, some such effects may be characterized by a very small amount of initial harm, followed by much larger benefit, but then generally ending in severe harm or lethality when the dose is excessively high.

Such tri-phasic dose-response behavior makes much more sense than just a biphasic inverse hormesis effect since it also considers the negative effects of excess dosages. It is entirely possible that all inverse hormesis dose-response curves may need to be at least tri-phasic, always ending in extreme dosages that result in terminal harm.

An example might be helpful, demonstrating the essential nature of a tri-phasic inverse hormesis for biological responses to both physical substance and energy dosing:

Seeds will often germinate when a small amount of moisture, warmth, and sunlight is available. While these small amounts may be enough to promote initial germination, they may not be adequate to sustain the growth of the plant. The observations we would expect for seed germination would look like this for various dosages of sunlight, warmth, and moisture:

  1. A little bit of each may induce germination, but cannot sustain growth (harmful)
  2. Just the right amount of each results in healthy plants (beneficial)
  3. Too much of any or all (sunlight, warmth, moisture) will harm or kill the plant (harmful)

This seems so simple that most children would know this intuitively, so much so that it is embedded in nursery stories (Goldilocks and the Three Bears). But this is far more nuanced than the commonly expressed scientific view of a typical toxicological dose-response; it is an example of tri-phasic inverse hormesis. It is intuitively clear and logical that this is a very reasonable biological dose-response characteristic. There is practical as well as scientific evidence that such responses do exist and may best characterize most biological responses if a sufficient range of dosages are studied, and this may help elucidate a very large portion of the often-contradictory data on the health effects of foods, supplements, external stimuli, and environmental exposures.

Extending this concept further, another insight may be that poly-phasic dose responses would generally be expected to end with an excessive dose that results in harm and death. Prior to death, larger doses are always possible. Even the most benign substances can result in overdose and even death, as evidenced by water intoxication, which is the result of severe electrolyte imbalance resulting from excessive water intake, which can lead to death. So we might expect that all poly-phasic dose-response curves must end at the highest doses in a negative phase, terminating in a state where further dosing is not possible.

Thus, the typical view of hormesis, beginning with benefit from very low doses but harm from larger doses, may typically be bi-phasic (good → bad), or may be poly-phasic with an even number of phases, always ending in doses at the highest level being observed to be harmful, toxic, or lethal. The converse would be true for inverse-hormesis-type dose responses, where the initial very low dose is harmful, followed by alternating phases of benefit and harm, with the final highest dose resulting in harm, toxicity, and eventually death. In this type of dose-response, the general case might be a tri-phasic response (bad → good → bad), with the possibility of a poly-phasic response, always expected to end in a harmful, toxic, or lethal dose, and therefore the phases of a poly-phasic inverse-hormesis response, beginning by definition with a harmful response, will always be odd in number. Once again, such multi-phasic behavior greater than two or three phases in dose-response curves may be exceedingly rare, or they may just not be often reported in the scientific literature simply because a sufficiently broad range of doses was not studied, both at the high- and low-end of the dosing range.

In the specific case of seed germination, these results are suggestive that PEMF may play a role in triggering certain events related to germination, and that inadequate triggering (dose is too low) may result in a failure of the seed to fully activate the mechanisms of germination. This may result from the partial activation of enzyme or other signaling systems in the seed which are adequate to initiate germination but insufficient to sustain it. Mechanistic hypotheses along these lines may contribute to the biophysical understanding of the observed effects of PEMF on seed germination.

The take away messages here are that the response of seeds to PEMF exposure is not simply linear with dose, and is probably not a typical sigmoidal dose-response curve. These findings remain to be repeated and verified (or not) independently before too much confidence should be invested in this unexpected result. But the observations overall open the possibility that the biological response to PEMF may be more generally biphasic or triphasic. This is very difficult to assess in most biological systems because a suitable biomarker for the effects of PEMF is not generally available, and even simple consistent benefits at a single dose of PEMF have been difficult to demonstrate. These measurements are often made much more difficult by the time periods involved (weeks, months, or sometimes years) and the qualitative nature of many of the most striking observed benefits of PEMF in humans. One of the principal reasons for this series of experiments was to use a very simple model that could be done in large numbers (hundreds to thousands of seeds), using a simple biomarker (seed germination: YES or NO and rate), with a reasonably short and consistent time frame (a few days or weeks). Thus, the observed inverse hormesis effect in this model system may be more generally applicable to the biologic responses to PEMF, but may not have been observed and reported in mammalian/human systems simply because typical mammalian system responses are too nuanced, complex, have a long-time frame, are more difficult to work with than plants and seeds, and lack an equivalently simple biomarker.

This result further points toward the possibility that PEMF responses may be bi-phasic, tri-phasic, or even poly-phasic with more than three distinct phases. It is also of importance to recognize that, for any of the dose-response curves that have been discussed, it is always possible that there is a threshold below which there will be no detectable effect, either positive or negative, and this should be factored into any considerations of the effects of PEMF on living systems.

Another point is that multi-phasic dose-response behavior may often be incorrectly reported as having fewer phases. For example, it may appear to have an initial phase that is “below threshold”, that is, it shows neither benefit nor harm, no effect positive or negative. It may of course be the case that sub-threshold responses are indeed below threshold and there is no biological response, or in some cases it may simply be that the initial phase is below the detection threshold due to measurement signal noise, or inadequate sensitivity or dynamic range of the measurement being employed, or may simply be due to the use of inadequately low doses.

Summary and Conclusions

High-germination-rate seeds such as lettuce appear to respond to PEMF when the correct wave form and dosage is used, resulting in even higher germination rates of 92% versus 79% for control, with slightly better than the threshold necessary for statistical significance at p ≤ 0.05.

Low-germination-rate seeds such as strawberry seeds may show a very slight improvement in germination rate when PEMF is applied, but this effect was very small for the PEMF type and dosage studied, and this possible effect of PEMF does not appear to interact with other common pre-planting procedures such as pre-freezing or pre-soaking of the seeds.

For intermediate-germination-rate seeds (bell pepper), PEMF using either 3 or 15 pulses per second has a very clear bi- or tri-phasic inverse hormesis effect on seed germination which appears to depend upon the number of PEMF pulses of specified waveform shape and intensity, and does not appear to be primarily an effect of “frequency” or time (duration) of exposure. Under the most favorable conditions studied (15 pps applied for 30 minutes, measured 3 weeks after planting), the germination rate was increased by 30% greater than baseline (70%) to 91%.

One over-all observation that does seem to be clear is that seeds from different plants respond quite differently to PEMF. This may be because different seeds have different dose-response characteristics with respect to PEMF, and the optimal PEMF parameters and doses may differ for seeds of different plants. Or it may be that different seeds may have different levels of sensitivity to stimulation by PEMF even at optimal stimulation levels. In any event, it is clear that any meaningful understanding of the effect of PEMF on plants and their seeds will depend upon (1) careful and detailed characterization of the PEMF pulse waveform and other key parameters, (2) determination of the PEMF dose-response characteristics for each type of seed over a wide range of dosages. This full spectrum dose-response characterization needs to be carried out for any seed type where PEMF is intended for use in a large-scale agricultural application.

Statistics and Raw Data

For lettuce seeds, statistics were carried out using a simple on-line statistics calculator:

For each of the tests involving a change in the germination rate between two groups, Z-scores were calculated () for two population proportions. Significance level was set to 0.05, and a two-tailed hypothesis was assumed.

Detailed statistical analysis of the strawberry seed germination is given in the Appendix.

Bell pepper germination raw data are available upon request.

For those who wish to use a different analysis, or who wish to check the calculations in this manuscript, copies of the original raw data are available upon request. Additional methodological detail is also freely available by contacting the senior and corresponding author (RG Dennis at: )

Statement of Potential Conflict of Interest

The primary author of this report (R.G. Dennis) declares both a scientific and a commercial interest in ICES®-PEMF technology: He is owner of Micro-Pulse LLC (manufacturer of the technology), holds several patents for ICES®-PEMF technology and receives royalty payments from NASA-Johnson Space Center for the commercial licensing of this technology, which he developed in its initial form (TVEMF) as a consultant for NASA in the mid-1990’s.

Appendix I: Analysis of Strawberry Germination Data

Data for WEEK 2 One Way Analysis of Variance

Data source: Data 1 in Notebook1

Normality Test (Shapiro-Wilk) Failed (P < 0.050)

Test execution ended by user request; ANOVA on Ranks begun

Kruskal-Wallis One Way Analysis of Variance on Ranks

Data source: Data 1 in Notebook1

Group N Missing Median 25% 75%
FSP 19 0 2 2 4
FS- 19 0 2 1 4
F-P 19 0 3 2 4
F-- 19 0 3 2 4
-SP 19 0 2 1 4
-S- 19 0 3 1 4
--P 19 0 5 4 6
--- 19 0 5 3 6
Table 6.H = 45.132 with 7 degrees of freedom. (P = <0.001)

The differences in the median values among the treatment groups are greater than would be expected by chance; there is a statistically significant difference (P = <0.001)

To isolate the group or groups that differ from the others use a multiple comparison procedure.

All Pairwise Multiple Comparison Procedures (Tukey Test):

Comparison Diff of Ranks q P<0.05
--P vs FS- 1317.000 6.863 Yes
--P vs FSP 1111.5 5.792 Yes
--P vs -S- 1089.000 5.675 Yes
--P vs -SP 1011.5 5.271 Yes
--P vs F-- 876.500 4.568 Yes
--P vs F-P 799.5 4.166 No
--P vs --- 107.000 0.558 Do Not Test
--- vs FS- 1210.000 6.306 Yes
--- vs FSP 1004.5 5.235 Yes
--- vs -S- 982.000 5.117 Yes
--- vs -SP 904.5 4.714 Yes
--- vs F-- 769.500 4.01 No
--- vs F-P 692.5 3.609 Do Not Test
F-P vs FS- 517.5 2.697 No
F-P vs FSP 312 1.626 Do Not Test
F-P vs -S- 289.5 1.509 Do Not Test
F-P vs -SP 212 1.105 Do Not Test
F-P vs F-- 77 0.401 Do Not Test
F-- vs FS- 440.5 2.296 Do Not Test
F-- vs FSP 235 1.225 Do Not Test
F-- vs -S- 212.5 1.107 Do Not Test
F-- vs -SP 135 0.704 Do Not Test
-SP vs FS- 305.500 1.592 Do Not Test
-SP vs FSP 100 0.521 Do Not Test
-SP vs -S- 77.500 0.404 Do Not Test
-S- vs FS- 228.000 1.188 Do Not Test
-S- vs FSP 22.5 0.117 Do Not Test
FSP vs FS- 205.5 1.071 Do Not Test
Table 7.

Note: The multiple comparisons on ranks do not include an adjustment for ties.

A result of "Do Not Test" occurs for a comparison when no significant difference is found between the two rank sums that enclose that comparison. For example, if you had four rank sums sorted in order, and found no significant difference between rank sums 4 vs. 2, then you would not test 4 vs. 3 and 3 vs. 2, but still test 4 vs. 1 and 3 vs. 1 (4 vs. 3 and 3 vs. 2 are enclosed by 4 vs. 2: 4 3 2 1). Note that not testing the enclosed rank sums is a procedural rule, and a result of Do Not Test should be treated as if there is no significant difference between the rank sums, even though one may appear to exist.

Data for WEEK 4 One Way Analysis of Variance

Data source: Data 1 in Notebook1

Normality Test (Shapiro-Wilk) Failed (P < 0.050)

Test execution ended by user request; ANOVA on Ranks begun

Kruskal-Wallis One Way Analysis of Variance on Ranks

Data source: Data 1 in Notebook1

Group N Missing Median 25% 75%
FSP 19 0 5 3 6
FS- 19 0 3 1 5
F-P 19 0 4 3 6
F-- 19 0 4 4 6
-SP 19 0 4 2 6
-S- 19 0 4 3 5
--P 19 0 6 5 7
--- 19 0 7 5 9
Table 8.H = 30.056 with 7 degrees of freedom. (P = <0.001)

The differences in the median values among the treatment groups are greater than would be expected by chance; there is a statistically significant difference (P = <0.001)

To isolate the group or groups that differ from the others use a multiple comparison procedure.

All Pairwise Multiple Comparison Procedures (Tukey Test):

Comparison Diff of Ranks q P<0.05
--- vs FS- 1167.000 6.082 Yes
--- vs -S- 996.500 5.193 Yes
--- vs -SP 953 4.966 Yes
--- vs FSP 728 3.794 No
--- vs F-P 695.5 3.624 Do Not Test
--- vs F-- 653.000 3.403 Do Not Test
--- vs --P 223 1.162 Do Not Test
--P vs FS- 944.000 4.919 Yes
--P vs -S- 773.500 4.031 No
--P vs -SP 730 3.804 Do Not Test
--P vs FSP 505 2.632 Do Not Test
--P vs F-P 472.5 2.462 Do Not Test
--P vs F-- 430.000 2.241 Do Not Test
F-- vs FS- 514 2.679 No
F-- vs -S- 343.5 1.79 Do Not Test
F-- vs -SP 300 1.563 Do Not Test
F-- vs FSP 75 0.391 Do Not Test
F-- vs F-P 42.5 0.221 Do Not Test
F-P vs FS- 471.5 2.457 Do Not Test
F-P vs -S- 301 1.569 Do Not Test
F-P vs -SP 257.5 1.342 Do Not Test
F-P vs FSP 32.5 0.169 Do Not Test
FSP vs FS- 439 2.288 Do Not Test
FSP vs -S- 268.5 1.399 Do Not Test
FSP vs -SP 225 1.173 Do Not Test
-SP vs FS- 214.000 1.115 Do Not Test
-SP vs -S- 43.500 0.227 Do Not Test
-S- vs FS- 170.500 0.889 Do Not Test
Table 9.

Note: The multiple comparisons on ranks do not include an adjustment for ties.

A result of "Do Not Test" occurs for a comparison when no significant difference is found between the two rank sums that enclose that comparison. For example, if you had four rank sums sorted in order, and found no significant difference between rank sums 4 vs. 2, then you would not test 4 vs. 3 and 3 vs. 2, but still test 4 vs. 1 and 3 vs. 1 (4 vs. 3 and 3 vs. 2 are enclosed by 4 vs. 2: 4 3 2 1). Note that not testing the enclosed rank sums is a procedural rule, and a result of Do Not Test should be treated as if there is no significant difference between the rank sums, even though one may appear to exist.

Data for WEEK 2 Three Way Analysis of Variance

Data source: Data 1 in Notebook1

Balanced Design (No Interactions)

Dependent Variable: Week 2

Normality Test (Shapiro-Wilk) Passed (P = 0.055)
Equal Variance Test: Passed (P = 1.000)
Source of Variation DF SS MS F P
Freeze 1 34.722 34.722 (+inf) <0.001
Soak 1 98.492 98.492 (+inf) <0.001
PEMF 1 3.463 3.463 (+inf) <0.001
Residual 1 0 0
Total 7 151.393 21.628
Table 10.

The difference in the mean values among the different levels of Freeze are greater than would be expected by chance after allowing for the effects of differences in Soak and PEMF. There is a statistically significant difference (P = <0.001). To isolate which group(s) differ from the others use a multiple comparison procedure.

The difference in the mean values among the different levels of Soak are greater than would be expected by chance after allowing for the effects of differences in Freeze and PEMF. There is a statistically significant difference (P = <0.001). To isolate which group(s) differ from the others use a multiple comparison procedure.

The difference in the mean values among the different levels of PEMF are greater than would be expected by chance after allowing for the effects of differences in Freeze and Soak. There is a statistically significant difference (P = <0.001). To isolate which group(s) differ from the others use a multiple comparison procedure.

Multiple Comparisons versus Control Group (Holm-Sidak method):

Overall significance level = 0.05

Comparisons for factor: Freeze

Comparison Diff of Means t P P<0.050
No Freeze vs. Freeze 4.167 (+inf) <0.001 Yes
Table 11.

Comparisons for factor: Soak

Comparison Diff of Means t P P<0.050
No Soak vs. Soak 7.018 (+inf) <0.001 Yes
Table 12.

Comparisons for factor: PEMF

Comparison Diff of Means t P P<0.050
no PEMF vs. PEMF 1.316 (+inf) <0.001 Yes
Table 13.

Power of performed test with alpha = 0.0500: for Freeze: 1.000

Power of performed test with alpha = 0.0500: for Soak: 1.000

Power of performed test with alpha = 0.0500: for PEMF: 1.000

Least square means for Freeze:

Group Mean
Freeze 11.732
No Freeze 15.899
Std Err of LS Mean = 0.000
Table 14.

Least square means for Soak:

Group Mean
Soak 10.307
No Soak 17.325
Std Err of LS Mean = 0.000
Table 15.

Least square means for PEMF:

Group Mean
PEMF 14.474
no PEMF 13.158
Std Err of LS Mean = 0.000
Table 16.

Data for WEEK 3 Three Way Analysis of Variance

Data source: Data 1 in Notebook1

Balanced Design (No Interactions)

Dependent Variable: Week 3

Normality Test (Shapiro-Wilk) Passed (P = 0.124)
Equal Variance Test: Passed (P = 1.000)
Source of Variation DF SS MS F P
Freeze 1 33.814 33.814 33.284 0.109
Soak 1 106.337 106.337 104.669 0.062
PEMF 1 0.00601 0.00601 0.00592 0.951
Residual 1 1.016 1.016
Total 7 173.605 24.801
Table 17.

The difference in the mean values among the different levels of Freeze are not great enough to exclude the possibility that the difference is just due to random sampling variability after allowing for the effects of differences in Soak and PEMF. There is not a statistically significant difference (P = 0.109).

The difference in the mean values among the different levels of Soak are not great enough to exclude the possibility that the difference is just due to random sampling variability after allowing for the effects of differences in Freeze and PEMF. There is not a statistically significant difference (P = 0.062).

The difference in the mean values among the different levels of PEMF are not great enough to exclude the possibility that the difference is just due to random sampling variability after allowing for the effects of differences in Freeze and Soak. There is not a statistically significant difference (P = 0.951).

Power of performed test with alpha = 0.0500: for Freeze: 0.341

Power of performed test with alpha = 0.0500: for Soak: 0.570

Power of performed test with alpha = 0.0500: for PEMF: 0.0926

Least square means for Freeze:

Group Mean
Freeze 16.886
No Freeze 20.998
Std Err of LS Mean = 0.504
Table 18.

Least square means for Soak:

Group Mean
Soak 15.296
No Soak 22.588
Std Err of LS Mean = 0.504
Table 19.

Least square means for PEMF:

Group Mean

Group Mean
PEMF 18.969
no PEMF 18.914
Std Err of LS Mean = 0.504
Table 20.

Data for WEEK 4 Three Way Analysis of Variance

Data source: Data 1 in Notebook1

Balanced Design (No Interactions)

Dependent Variable: Week 4

Normality Test (Shapiro-Wilk) Failed (P < 0.050)
Equal Variance Test: Passed (P = 1.000)
Source of Variation DF SS MS F P
Freeze 1 16.255 16.255 169 0.049
Soak 1 98.492 98.492 1024 0.02
PEMF 1 0.866 0.866 9 0.205
Residual 1 0.0962 0.0962
Total 7 156.202 22.315
Table 21.

The difference in the mean values among the different levels of Freeze are greater than would be expected by chance after allowing for the effects of differences in Soak and PEMF. There is a statistically significant difference (P = 0.049). To isolate which group(s) differ from the others use a multiple comparison procedure.

The difference in the mean values among the different levels of Soak are greater than would be expected by chance after allowing for the effects of differences in Freeze and PEMF. There is a statistically significant difference (P = 0.020). To isolate which group(s) differ from the others use a multiple comparison procedure.

The difference in the mean values among the different levels of PEMF are not great enough to exclude the possibility that the difference is just due to random sampling variability after allowing for the effects of differences in Freeze and Soak. There is not a statistically significant difference (P = 0.205).

All Pairwise Multiple Comparison Procedures (Holm-Sidak method):

Overall significance level = 0.05

Comparisons for factor: Freeze

Comparison Diff of Means t P P<0.050
PEMF vs. no PEMF 2.851 13.000 0.049 Yes
Table 22.

Comparisons for factor: Soak

Comparison Diff of Means t P P<0.050
PEMF vs. no PEMF 7.018 32.000 0.020 Yes
Table 23.

Comparisons for factor: PEMF

Comparison Diff of Means t P P<0.050
PEMF vs. no PEMF 0.658 3.000 0.205 No
Table 24.

Power of performed test with alpha = 0.0500: for Freeze: 0.690

Power of performed test with alpha = 0.0500: for Soak: 0.988

Power of performed test with alpha = 0.0500: for PEMF: 0.194

Least square means for Freeze:

Group Mean
Freeze 18.75
No Freeze 21.601
Std Err of LS Mean = 0.155
Table 25.

Least square means for Soak:

Group Mean
Soak 16.667
No Soak 23.684
Std Err of LS Mean = 0.155
Table 26.

Least square means for PEMF:

Group Mean
PEMF 20.504
no PEMF 19.846
Std Err of LS Mean = 0.155
Table 27.

Da ta for WEEK 2 vs. WEEK 3 vs. Week 4 One Way Analysis of Variance

Data source: Data 1 in strawberry data

Normality Test (Shapiro-Wilk) Failed (P < 0.050)

Test execution ended by user request; ANOVA on Ranks begun

Kruskal-Wallis One Way Analysis of Variance on Ranks

Data source: Data 1 in strawberry data

Group N Missing Median 25% 75%
Week 2 8 0 12.171 10.581 19.024
Week 3 8 0 17.215 15.789 23.026
Week 4 8 0 19.627 16.502 23.794
Table 28.

The differences in the median values among the treatment groups are greater than would be expected by chance; there is a statistically significant difference (P = 0.039)

To isolate the group or groups that differ from the others use a multiple comparison procedure.

All Pairwise Multiple Comparison Procedures (Tukey Test):

Comparison Diff of Ranks q P<0.05
Week 4 vs Week 2 70 3.5 Yes
Week 4 vs Week 3 20 1 No
Week 3 vs Week 2 50 2.5 No
Table 29.

Note: The multiple comparisons on ranks do not include an adjustment for ties.


  1. Bassett C. A. L., Becker R. O.. Generation of Electric Potentials by Bone in Response to Mechanical Stress. Science. 1962; 137(3535)DOI
  2. Andrew C., Bassett L., Pawluk R. J., Pilla A. A.. Augmentation of Bone Repair by Inductively Coupled Electromagnetic Fields. Science. 1974; 184(4136)DOI
  3. Dennis Robert, Dennis John. Reversal of Renal Insufficiency in an Aging Cat: A 5-year Multi-Crossover Case Study. The Journal of Science and Medicine. 2020; 2(2)DOI
  4. Fenn JE. Effect of pulsed electromagnetic energy (Diapulse) on experimental hematomas. Can Med Assoc J. 2969; 100(5):251-254.
  5. Hubbard Devin. Electroceutical Technology: Anti-Inflammatory Effects Of 40-160 T/S Inductively Coupled Electrical Stimulation (ICES) In The Acute Inflammation Model. The Journal of Science and Medicine. 2020; 2(2)DOI
  6. Dennis Robert. Inductively Coupled Electrical Stimulation - Part I: Overview and First Observations. The Journal of Science and Medicine. 2019; 1(1)DOI
  7. Dennis Robert. Inductively Coupled Electrical Stimulation - Part 2: Optimization of parameters for orthopedic injuries and pain. The Journal of Science and Medicine. 2020; 1(2)DOI
  8. Dennis Robert. Inductively Coupled Electrical Stimulation - Part 3: PEMF Systems for use in Basic Research with Laboratory Animals and In Vitro. The Journal of Science and Medicine. 2020; 2(3)DOI
  9. Dennis R. Letter to the Editor: Strawberries: Facts, Truth, Misinformation, Publication Bias and the Importance of Negative Results. The Journal of Science and Medicine. 2021; 2(4)DOI
  10. Pietruszewski Stanisław, Martínez Elvira. Magnetic field as a method of improving the quality of sowing material: a review. International Agrophysics. 2015; 29(3)DOI
  11. Reinke. Untersuchungen der Wachstum. Studies of Growth. 1876; Botan. Ztg
  12. KATSENIOS Nikolaos, KAVVADIAS Victor, THEOCHAROPOULOS Sideris, BILALIS Dimitrios, IOANNOU Zaharias, PAPADOPOULOS Antonios, LIAKOPOULOU Nektaria. Influence of Pulsed Electromagnetic Field on Plant Growth, Nutrient Absorption and Yield of Durum Wheat. Notulae Scientia Biologicae. 2015; 7(4)DOI
  13. Efthimiadou Aspasia, Katsenios Nikolaos, Karkanis Anestis, Papastylianou Panayiota, Triantafyllidis Vassilios, Travlos Ilias, Bilalis Dimitrios J.. Effects of Presowing Pulsed Electromagnetic Treatment of Tomato Seed on Growth, Yield, and Lycopene Content. The Scientific World Journal. 2014; 2014DOI
  14. Aleman EI, Mbogholi A, Boix Y, Gonzalez-Ohnedo J, Chalfun A. Effects of EMFs on some biological parameters in coffee plants (Coffea arabica L.) obtained by in vitro propagation.. Polish J. Environ. Stud.. 23(2014):95-101.
  15. Izmailov Andrey, Smirnov Igor, Khort Dmitriy, Filippov Rostislav, Kutyrev Alexey. Magnetic-pulse processing of seeds of berry crops. Research in Agricultural Engineering. 2018; 64(No. 4)DOI
  16. Xu Chunxiao, Lv Yan, Chen Chuanfang, Zhang Yuxia, Wei Shufeng. Blue light-dependent phosphorylations of cryptochromes are affected by magnetic fields in Arabidopsis. Advances in Space Research. 2014; 53(7)DOI
  17. Yan Xiaofei, Wang Zhongyi, Huang Lan, Wang Cheng, Hou Ruifeng, Xu Zhilong, Qiao Xiaojun. Research progress on electrical signals in higher plants. Progress in Natural Science. 2009; 19(5)DOI
  18. Borgman HH. De invloed van een koudebehandeling op de kieming van aardbeizaad. Meded. Dir. Tuinb. 1950; 13:13-15.
  19. Scott DH, Ink DP. Treatments to hasten the emergence of seedlings of blueberry and strawberry. Proc. Amer. Soc. Hort. Sci. 1955; 66:237-242.
  20. Bringhurst ES, Voth V. Effect of stratification on strawberry seed germination. Proc. Amer. Soc. Hort. Sci. 1957; 70:144-149.
  21. Jonkers H. Accelerated flowering of strawberry seedlings. Euphytica. 1958; 7(1):41-46.
  22. Adam JI, Wilson D. Factors affecting the germination of strawberry seed.. 1967.
  23. Thompson P.A.. The use of chilling and chemical treatments to promote rapid germination of strawberry achenes. Journal of Horticultural Science. 1969; 44(2)DOI
  24. Wilson D., Goodall Anne, Reeves Jane. An improved technique for the germination of strawberry seeds. Euphytica. 1973; 22(2)DOI
  25. Guttridge C.G., Bright S. Accelerating and synchronizing germination of strawberry seeds by osmotic pre-treatments. Euphytica. 1978; 27(3)DOI
  26. The importance of no evidence. Nature Human Behaviour. 2019; 3(3)DOI
  27. Nissen Silas Boye, Magidson Tali, Gross Kevin, Bergstrom Carl T. Publication bias and the canonization of false facts. eLife. 2016; 5DOI
  28. Fanelli Daniele. Negative results are disappearing from most disciplines and countries. Scientometrics. 2011; 90(3)DOI
  29. Turner Erick H., Matthews Annette M., Linardatos Eftihia, Tell Robert A., Rosenthal Robert. Selective Publication of Antidepressant Trials and Its Influence on Apparent Efficacy. New England Journal of Medicine. 2008; 358(3)DOI
  30. Figure 13:, retrieved January, 2021.
  31. Mattson Mark P.. Hormesis defined. Ageing Research Reviews. 2008; 7(1)DOI
  32. CALABRESE E, BLAIN R. The occurrence of hormetic dose responses in the toxicological literature, the hormesis database: an overview. Toxicology and Applied Pharmacology. 2005; 202(3)DOI
  33. Pearce O. M. T., Laubli H., Verhagen A., Secrest P., Zhang J., Varki N. M., Crocker P. R., Bui J. D., Varki A.. Inverse hormesis of cancer growth mediated by narrow ranges of tumor-directed antibodies. Proceedings of the National Academy of Sciences. 2014; 111(16)DOI
  34. Paracelsus. dritte defensio. 1538.
  35. Kong Eva Yi, Cheng Shuk Han, Yu Kwan Ngok. Biphasic and triphasic dose responses in zebrafish embryos to low-dose 150 kV X-rays with different levels of hardness. Journal of Radiation Research. 2016; 57(4)DOI
  36. CHOI Viann Wing Yan, YUM Emily Hoi Wa, KONISHI Teruaki, OIKAWA Masakazu, CHENG Shuk Han, YU Kwan Ngok. Triphasic Low-dose Response in Zebrafish Embryos Irradiated by Microbeam Protons. Journal of Radiation Research. 2012. DOI