Table 4 also shows that in the native bracken soils, total 

 percentage of germination is not significantly different 

 than germination on potting mix, with the exception of 

 Douglas-fir. When total germination between the two soil 

 types is the same, differences in percentage of seed ger- 

 minating but dying can be attributed to soil type. Said 

 another way, stratification of seed in bracken environ- 

 ments did not reduce germination, but major differences 

 in survival developed after seed germination. Douglas-fir 

 germination seems to have been affected by stratification 

 because seed overwintered on bracken soil has significant- 

 ly lower total percentage germination than on potting mix 

 transported to the same site. 



Engelmann spruce had very low germination on all 

 treatment and soil types. Because 32.9 percent of spruce 

 seed germinated in the stratified control, we believe most 

 spruce seed was killed (or germination energy reduced) 

 during the overwintering period. 



Results from the forested site are inadequate as a con- 

 trol for this study. Conditions under the dense tree over- 

 story are quite different than in the open bracken glades. 

 Fewer seeds germinated in the forested treatment than in 

 any other treatment. Because conifer seed germinated and 

 survived on forest soil in our greenhouse bioassay tests, 

 we feel the loss of seed viability is associated with condi- 

 tions in the understory and not an inherent problem with 

 the soil. 



CONCLUSIONS AND DISCUSSION 



The review of literature showed that bracken fern is a 

 troublesome weed worldwide. It can quickly invade 

 disturbed forests and grazed lands. Increases in bracken 

 occurrence are attributable to human activities. Bracken is 

 well adapted to maintaining dominance because of its un- 

 palatability to browsing, chemical defenses against insects 

 (Cooper-Driver 1976), ability to sprout following disturb- 

 ance, and production of phytotoxins. 



Bracken fern phytotoxins could reduce or eliminate 

 other species in several ways, such as killing seed during 

 stratification, reducing seed germination energy, reduced 

 vigor that predisposes seedhngs to other causes of mor- 

 tality, or the death of plants caused by absorption of 

 phytotoxins. Our studies on bracken fern were conducted 

 to test the allelopathic potential of bracken fern in north- 

 ern Idaho and to elucidate mechanism(s) that account for 

 bracken dominance. 



Tests conducted at our Moscow, ID, greenhouse found 

 some statistically significant reductions in germination and 

 growth attributable to bracken-dominated soil. Tests at a 

 bracken fern glade showed even more striking results in 

 that most seed died when germinated on soil dominated by 

 bracken. The large percentage of seedlings that died helps 

 explain the nearly total exclusion of conifers from bracken 

 glades in northern Idaho. 



Based on the results of this research, we feel an impor- 

 tant mechanism limiting establishment of conifers in 

 bracken glades of northern Idaho is death of germinants 

 as the radicle penetrates the upper surface of the soil. 

 Death probably results when the germinant absorbs phyto- 

 toxins that have accumulated in the soil over a number of 

 years. ^ 



We hypothesize that phytotoxins are concentrated near 

 the soil surface. This follows from finding reduced ger- 

 mination and growth in the greenhouse bioassay tests only 

 in tests having bracken glade soil from the 0- to 5-cm 

 depth. Even for the 0- to 5-cm depth, more seed survived 

 in the greenhouse test compared to seed germinated at 

 the study site. Mixing of the soil in the 0- to 5-cm depth 

 may have diluted the phyto toxin, indicating that if a 

 phytotoxic layer is present, it is thin. In addition, radicle 

 lengths of dead seed in the seed stratification study were 

 less than 3 cm. 



Other investigators have reported that allelopathic ef- 

 fects are greatest near the soil surface. Gliessman and 

 Muller (1978) collected raindrip from bracken fronds under 

 field conditions and showed allelopathic effects using 

 bioassay tests. They hypothesized that toxins were being 

 concentrated in the upper soil layers where they remain 

 even with heavy rainfall. Horsley (1977) thought that tox- 

 ins from three species he investigated may be bound to 

 the soil complex and at least 1 year is necessary for toxic 

 plant residues to be neutralized, destroyed, or leached. 

 Muller (1966), working in California chaparral, found that 

 herbs were able to become estabhshed following a fire or 

 after a layer of soil 5 cm thick was removed. 



Discussion of bracken fern communities in northern 

 Idaho would be incomplete without considering inter- 

 actions with pocket gophers. The role of gophers in 

 bracken communities is seldom mentioned in the litera- 

 ture. Gophers feed on forbs, grasses, shrubs, conifers, 

 seeds, and just about any vegetation not poisonous 

 (Teipner and others 1983). We do not know if pocket 

 gophers feed on bracken, but few other species grow in 

 these bracken glades where gopher populations are high. 



Gophers deposit subsurface soil above ground in mounds 

 and casts. This buries existing vegetation and, over time, 

 would create a mosaic of microsites differing in succes- 

 sional status. These microsites could account for the 

 presence of forbs, grasses, and sedges which, in turn, 

 supply food for the gophers. These species could become 

 established on fresh mounds, eventually dying out as 

 bracken phytotoxins again accumulate. 



Shrubs and conifers could also become established on 

 gopher mounds. Tevis (1956) noted that an abundant red 

 fir (Abies magnifica) cone crop in northern California 

 resulted in good reproduction on ground bared by pocket 

 gophers. (Interestingly, he noted that gopher-bared ground 

 in nonbracken areas had "abundant" seedlings whereas 

 seedlings on gopher-bared ground in bracken areas were 

 "moderately common.") Seed predation by gophers and 

 mice could account for loss of some shrub and conifer 

 seeds, but this is unlikely to be the whole story. 



An additional explanation for the lack of woody species 

 in northern Idaho bracken glades concerns decomposition 

 of bracken fronds under the snow. Even the casual 

 observer is impressed with the large amount of bracken 

 frond biomass at the Eagle Point study site. Znerold 

 (1979) calculated frond density at 286,300 stems/ha and 

 frond ovendry weight at 3,900 kg/ha. Each autumn fronds 

 die, fall over, and are pressed to the earth by snow. In the 

 spring, little evidence of old bracken fronds exists except 

 for portions of the thick petioles that are not totally 

 decomposed. A great deal of decomposition must take 



9 



