Posted on

weed and seed program wiki

General Description

The following description of Rubus discolor is taken from Munz and Keck (1973) [1] .

Rubus discolor is a robust, sprawling, more or less evergreen, glandless shrub of the Rose Family (Rosaceae). The shrubs appear as “great mounds or banks” [2] , with some of the canes standing up to 3 m tall. Other canes are decumbent, trailing or scandent up to 20-40 feet long [3] , frequently taking root at the tips. The primocanes are pilose-pubescent, becoming nearly glabrous with age. These are very strongly angled and furrowed, bearing well-spaced, heavy, broad-based, straight or somewhat curved prickles 6-10 mm long. Primocane leaves are 5-foliolate, glabrous above when mature and cano-pubescent to cano-tomentose beneath. There are hooked prickles on the petioles and petiolules. The leaflets are large and broad with the terminal leaflet roundish to broad oblong. Leaflets are abruptly narrowed at the apex, unequally and coarsely serrate-dentate. Floricane leaflets are 3-5 foliolate and smaller than on the primocanes.

The inflorescence is a large terminal cluster with branches in the lower axils. The peduncles and pedicels are cano-tomentose and prickly. The flowers are white or rose colored, 2-2.5 cm across, with broad petals. Sepals are broad, cano-tomentose, conspicuously pointed and soon reflexed, approximately 7-8 mm long. The roundish fruit is black and shiny, up to 2 cm long, with large succulent drupelets. The fruit ripens late compared with native blackberries and over a considerable interval [2] , from midsummer to autumn. [3]



Contrary to its common name, Himalaya-berry is a native of western Europe. [1] There is no botanical evidence to show that it is native of the Himalayan region. It may have found its way there as a cultivar. [3] Rubus discolor was probably first introduced to North America in 1885 as a cultivated crop as well. [2] By 1945 R. discolor had become naturalized along the West Coast. It also occurred in nursery and experiment grounds along the East Coast and in Ohio [2] by this time. R. discolor occurs mainly in areas with an average annual rainfall greater than 76 cm, at altitudes up to 1800 m, and on both acidic and alkaline soils. [4] It forms impenetrable thickets in wastelands, pastures, and forest plantations. [5] It grows along roadsides, creek gullies, river flats, fence lines [6] , and right-of-way corridors. R. discolor tends to prefer wet sites even in relatively wet climates. [7]


Rubus discolor is a part of the Rubus fruticosus aggregate.

VEGETATIVE REPRODUCTION Rooting at Cane Tips: R. discolor can form roots at cane apices. Amor (1974a) [8] observed canes growing to a height of 40 cm before they arched over and trailed on the ground. Daughter plants developed where these canes rooted, forming only on first-year canes. All canes produced berries in the second year and then died, senescence commencing near the middle and at the apices of canes without daughter plants. Re-entry of canes into the center of the thicket resulted in an impenetrable mass of prickly canes within 2-1/2 years. Individual canes may only live 2-3 years, yet reach a density of 525 canes per square meter. A large quantity of litter and standing dead canes develops in old thickets. [4]

Canes of R. discolor can grow to lengths of up to 7 m in a single season. At one site observed by Amor (1974a) [8] , the mean horizontal projection of 50 first-year canes was 3.3 m. Ninety-six percent of these canes had daughter plants at their apices. Lateral branches on some canes had also formed daughter plants.

Adventitious Shoots on Lateral Roots: The root crown on R. discolor can be up to 20 cm in diameter, from which many lateral roots grow at various angles. One measured root had a maximum depth of 90 cm but was more than 10 m long. [9] Adventitious shoots (suckers) are occasionally formed on the roots and may emerge from a depth of 45 cm. Blackberries also readily propagate from root pieces and cane cuttings. [8] In less than two years a cane cutting can produce a thicket 5 m in diameter. [5]

SEXUAL REPRODUCTION Seed Production: Himalaya-berry thickets can produce 7000-13,000 seeds per square meter. [8] When grown in dense shade, however, most species of blackberry do not form seeds. Good seed crops occur nearly every year. [10]

Each seed contains two ovules, but one usually aborts. [11] The seeds are contained in berries, which consist of a number of loosely adhering drupelets. Berries ripen and turn black during the summer on canes more than one year old. Each drupelet contains one seed.

Seed Dispersal: Dispersal of Rubus seeds by birds has been reported by several authors. [8] Passage of the seed through the digestive tract of birds may improve germination. Dispersal may also be accomplished by omnivorous mammals such as foxes, as suggested by numerous authors with respect to Australia. [12] [7] The prompt invasion of cutover lands by Rubus indicates that the dispersed seeds remain viable in the soil for several years. [10]

Germination and Seedling Establishment: Blackberry seeds germinate mainly in the spring, but there is little germination of seed in the first spring after the seed is formed. [11] Brinkman (1974) [10] reported 33 percent germination in un-aged R. discolor seeds. A three year field trial showed only 10 percent germination in Rubus discolor. [4]

In Australia R. discolor seedlings receiving less than 44 percent of full sunlight did not survive. [8] The slow growth of seedlings and their susceptibility to shading suggest that few seedlings would be expected to survive in dense pastures or forest plantations. Blackberry thickets are also poor sites for seedling development. Amor (1972) [4] counted less than 0.4 seedlings per square meter near thickets. The establishment of R. discolor seedlings depends on the availability of open habitats such as land neglected after cultivation, degraded pastures, and eroded soils along streams. [8] Although seedlings show the potential for rapid growth under laboratory conditions, they grow much slower in the field and are easily surpassed by the more rapid growth rate of daughter plants.

Rubus discolor colonizes areas initially disturbed and then neglected by humans. It is a perennial weed which is difficult to control due to its ability to regenerate from sections of root stock. The production of dense thickets, especially in wet areas, may hinder medium- to large-ized mammals in gaining access to water. Himalaya-berry may also displace native plant species. Seeds can be widely dispersed by berry-eating birds.

Himalaya-berry occurs on TNC’s Santa Rosa Plateau and McCloud River preserves in California.


Management Requirements

With proper management, areas infested with R. discolor can be restored to more desirable vegetation. Mechanical removal or burning may be the most effective ways of removing the mature plants. Subsequent treatment with herbicides should be conducted cautiously for two reasons: (1) R. discolor often grows in riparian areas and the herbicide may be distributed to unforeseen locations by running water, and (2) some herbicides promote vegetative growth from lateral roots. Himalaya-berry re-establishment may be prevented by planting fast-growing shrubs or trees, since the species is usually intolerant of shade. Regrowth has also been controlled by grazing sheep and goats in areas where the mature plants have been removed.

Monitoring is needed to determine the effectiveness of management practices.

Detailed observations focused on the vegetational change of the affected area over time will help to determine what method of control would be most efficient.

No quantitative monitoring studies of Himalaya-berry were discovered in this research. Since it is not considered a major agricultural weed in California, there has been little interest or funding available for detailed sampling programs. Monitoring efforts so far have been only qualitative in nature: Has it invaded a site? Does it re-establish itself following control treatment?

Weed control involves three fundamental objectives: prevention, eradication and control.

From a practical viewpoint, methods of weed management are commonly categorized under the following categories: physical, managerial, biological and chemical controls, and prescribed burning. [13] Physical methods include both manual and mechanical means. Managerial methods include the encouragement of competitive displacement by native plants and prescribed grazing. Biological control is usually interpreted as the introduction of insects or pathogens which are highly selective on a particular weed species. Chemical control includes both broadcast and spot application. Prescribed burning includes both broadcast burning and spot treatment with a flame thrower.

The most desirable approach is that of an integrated pest management plan. This involves the optimal integration of appropriate control strategies to control weeds. This approach is generally accepted as the most effective, economical and environmentally sound, long-term pest control strategy. [13] In cases where more than one control technique is used, the various techniques should be compatible with one another. Broadcast herbicide application, for example, may not work well with certain managerial techniques (i.e., plant competition).

See also  blue tahoe seeds


The physical control methods discussed below, manual and mechanical, produce slash (i.e., cutting debris) that can be disposed of by several techniques. If cut before seeds are produced it may be piled and left for enhancement of wildlife habitat (i.e., cover for small mammals). Debris may be fed through a mechanical chipper and used as mulch during revegetation procedures. Particularly with Rubus, care should be taken to prevent vegetative reproduction from cuttings. Burning slash piles is also an effective method of disposal.


Manual methods use hand labor to remove undesirable vegetation. These methods are highly selective and permit weeds to be removed without damage to surrounding native vegetation.

The Bradley Method is one sensible approach to manual control of weeds. [14] This method consists of hand weeding selected small areas of infestation in a specific sequence, starting with the best stands of native vegetation (those with the least extent of weed infestation) and working towards those stands with the worst weed infestation. Initially, weeds that occur singly or in small groups should be eliminated from the extreme edges of the infestation. The next areas to work on are those with a ratio of at least two natives to every weed. As the native plants stabilize in each cleared area, work deeper into the center of the most dense weed patches. This method has great promise on nature reserves with low budgets and sensitive plant populations. More detailed information is contained in Fuller and Barbe (1985) [14] .

Hand Pulling: This method may be used to destroy seedlings and young plants up to 1 m tall. Seedlings are best pulled after a rain when the soil is loose. This facilitates removal of the rooting system, which may resprout if left in the ground. Plants should be pulled as soon as they are large enough to grasp but before they produce seeds.

Hand Hoeing: Plants can be destroyed readily while they are still small by hand hoeing, either by cutting off their tops or by stirring the surface soil so as to expose the seedlings to the drying action of the sun. The object of hoeing is to cut off weeds without going too deeply into the ground and doing damage to the roots of desirable vegetation.

For plants up to 4 m tall a claw mattock is effective for removing the root crowns. The dirt around the root is loosened by the claw, and the plant is pulled out in the same way that a claw hammer is used to pull out nails.

Cutting: Manually operated tools such as brush cutters, power saws, axes, machetes, loppers and clippers can be used to cut R. discolor. This is an important step before many other methods are tried, as it removes the above-ground portion of the plant. In addition, for a thickly growing, multi-stemmed shrub such as R. discolor, access to the base of the shrub may not only be difficult but dangerous where footing is uncertain.

An advantage of cane removal over foliage herbicides is that cane removal does not stimulate sucker formation on lateral roots. Amor (1974b) [15] provides evidence that herbicides such as picloram and 2,4,5-T are not considerably more effective than cane removal. However, removal of canes alone is insufficient to adequately control Rubus discolor, as the root crown will simply resprout and produce more canes.

Hand Digging: The removal of rootstocks by hand digging is a slow but sure way of destroying R. discolor, a weed which resprouts from its roots. The work must be thorough to be effective as every piece of root that breaks off and remains in the soil may produce a new plant. Such a technique is only suitable for small infestations and around trees and shrubs where other methods are not practical.


Mechanical methods use mechanized equipment to remove above-ground vegetation. These methods are often non-selective in that all vegetation on a treated site is affected. Mechanical control is highly effective at controlling woody vegetation on gentle topography with few site obstacles such as root stumps or logs. Most mechanical equipment is not safe to operate on slopes over 30 percent. It is also of limited use where soils are highly susceptible to compaction or erosion or where excessive soil moisture is present.

Chopping, Cutting or Mowing: Rubus discolor plants may be trimmed back by tractor-mounted mowers on even ground or by scythes on rough or stony ground. Unwanted vegetation can be removed faster and more economically in these ways than by manual means and with less soil disturbance than with scarification. However, these methods are non-selective weed eradication techniques. They reduce the potential for biological control through plant competition and may open up new niches for undesirable vegetation. Wildlife forage is eliminated, cutting down on the general habitat value of the area. Another disadvantage of cutting, chopping or mowing is that perennial weeds such as Himalaya-berry usually require several cuttings before the underground parts exhaust their reserve food supply. If only a single cutting can be made, the best time is when the plants begin to flower. At this stage the reserve food supply in the roots has been nearly exhausted, and new seeds have not yet been produced. After cutting or chopping with mechanical equipment, Rubus may resprout from root crowns in greater density if not treated with herbicides.


Broadcast Burning: Large areas of weed infestation may be burned in order to remove the standing mature plants. This may be accomplished with a pre-spray of herbicides, to kill and desiccate plants, or without such spraying for notably flammable species. Used alone, this method will not prevent resprouting from root crowns. Burning is best followed by (1) stump herbicide treatment, (2) subsequent burning to exhaust soil seed bank and underground food reserves, and/or (3) revegetation with fast-growing native species. Other considerations for the use of prescribed burning include the time and cost of coordinating a burn and the soil disturbance resulting from firebreak construction.


Biological Competition: Sowing native plant species which have the potential to out-compete weedy exotics for important resources is usually a preventive method of weed control. In some cases later successional plants may be encouraged to take root among the unwanted vegetation.

In most cases Himalaya-berry prevents the establishment of other native plants and must be initially removed. Following physical removal or burning of mature plants, root crowns must be treated to prevent resprouting. Seedlings of native plant species usually cannot establish fast enough to compete with sprout growth from untreated stumps.

Some plant species inhibit the establishment or growth of other plants through the effects of allelopathy (i.e., biochemical interference by metabolic products). Native species with such properties may be propagated in treated areas to control re-establishment of blackberry, but as allelopathy is occasionally a trait of noxious weeds, it is wise not to replace an old problem species with a new one.

Prescribed Grazing: Amor (1974a) [8] described the effects of grazing by various animals on Himalaya-berry as follows: In an ungrazed area, 96% of the plants produced daughter plants; in areas grazed lightly by horses the number dropped to 11%; in areas grazed by cattle only 1% of all plants had daughter plants; and no plants had daughter plants in areas grazed by sheep.

In New Zealand the recognized method of Rubus control in the past has been the farming of large numbers of goats. This method has been effective in preventing canes from totally covering large areas. [16] [17] Crouchley (1980) [18] mentions that blackberry is readily eaten by goats throughout the year, even when there is an abundant supply of pasture and other plants.

In many areas of California the use of Angora and Spanish goats is showing promise as an effective control for Himalaya-berry. [19] In Cleveland National Forest goats are herded for firebreak management of brush species on over 79,000 acres of land. Goats are less costly to utilize than mechanical and chemical control methods. They can negotiate slopes too steep to manage with machines and do not pose the environmental dangers inherent with herbicides. [20]

A pioneer in the use of goats for weed control in urban settings is Richard Otterstad, owner of Otterstad’s Brush Clearing Service (718 Adams St., Albany, CA 94706, (415) 524-4063). The primary weed control “tools” utilized by Otterstad’s company are Angora goats and light-weight flexible fencing reinforced with electrified wire. Angoras are preferred over Spanish goats because their smaller size makes them easier to transport (Otterstad uses a pickup truck). Dairy goats were abandoned when Otterstad found them to be “goof-offs” when it came to eating. [19]

See also  pancake ice seeds

Goats prefer woody vegetation over most grasses or forbs, although Angoras have a higher tolerance for non-woody species. Since goats will trample or browse virtually any vegetation within a fenced area, any desirable trees or shrubs must be protected. Experience has shown that goats are most cost-effective when used to clear or suppress brush regrowth of one to four years old rather than to do initial clearing of dense tall, mature stands of vegetation. When faced with mature brush, goats will defoliate twigs and strip off bark but will leave standing the plant’s main superstructure which is too old and tough to tempt them. Sheep are more selective than goats in their food choices but function well in grazing down a variety of plants. Thus sheep grazing may be a practical alternative to mowing. It is important to properly manage sheep grazing due to soil compaction problems if sheep are allowed to graze an overly damp area. Sheep are valuable not only for weed control but also for their contribution of fertilizer to the soil and additional income from the sale of their wool. However, exotic seeds may be initially introduced from sheep droppings.

Chickens, surprisingly enough, are known to effectively digest (and destroy) all weed seeds passing through their crops and can thoroughly graze back vegetation in areas of up to one acre in size. Releasing chickens into an area after the mature plants are removed allows them to scratch and peck out weed seeds and potentially reduce the weed seed bank in the soil. [20]


The USDA will not support the introduction of herbivorous insects to control Himalaya-berry due to the risk these insects may pose to commercially important Rubus species. Please notify the California Field Office of The Nature Conservancy of any field observations in which a native insect or pathogen is seen to have detrimental effects on Rubus species. These reports will be used to update this Element Stewardship Abstract. Management techniques which may encourage the spread of species-specific agents may be desirable in controlling Himalaya-berry.


Detailed information on herbicides is available in such publications as Weed Science Society of America (1983) [21] or USDA (1984) [22] and will not be comprehensively covered here. Publications such as these give specific information on nomenclature, chemical and physical properties of the pure chemical, use recommendations and precautions, physiological and biochemical behavior, behavior in or on soils, and toxological properties for several hundred chemicals.

Herbicides may be applied non-selectively (i.e., broadcast applications) or selectively (i.e., spot treatments). Both types of treatments have advantages and disadvantages and will be discussed separately.

Broadcast Herbicide Application: Broadcast application of herbicides has become the mainstay of most weed control efforts today. This may be due to the illusion that it is a “quick fix” method of eradicating undesirable vegetation. Most herbicides so applied are non-selective and will kill most, if not all, of the vegetation sprayed. Those species which survive the treatment may, after repeated sprayings, form an herbicide-resistant vegetation cover, thus creating a more difficult problem to deal with. Such broadcast spraying may also kill off native plants which have the ability to out-compete exotic weeds.

Herbicides should be applied only when the plants are in full leaf. Results are poor if the plants are sprayed prior to this stage. The best results occur when plants are sprayed after seed-set. [23]

Broadcast herbicide application may be most effective where the weed infestation is very dense and needs to be killed and desiccated prior to burning. It may also be useful following the removal of mature plants so as to reduce the soil seed bank.

Picloram (Tordon ® ) is effective [24] [25] [26] [15] , but one application may not always be sufficient. [26] [27] Picloram suppresses cane regrowth but stimulates the development of adventitious shoots. [15] Foliage spraying is more effective in the summer than winter. [27]

Many other herbicides have been used in an attempt to control Himalaya-berry with varying degrees of effectiveness. Fosamine is not as effective for killing Rubus discolor, but it is more effective in controlling regrowth. [28] Neither Fosamine nor glyphosate [29] [30] provide long-term control of R. discolor. Blackberry control has also been accomplished with dicamba [31] , aminotriazole [27] , amitrole-thiocyanate [32] [4] and triclopyr ester. [33]

Spot Chemical Methods: Spot chemical methods consist of various techniques for manually applying herbicides to individual plants or small clumps of plants (such as stump resprouts). These methods are highly selective as only specific plants are treated. They are most efficient when the density of stems to be treated is low. In applying herbicides it is recommended that a dye be used in the chemical mixture to mark the treated plants and thus minimize waste.

Jones and Stokes Associates (1984) [34] reviewed a variety of spot chemical techniques. The following is an excerpt from this report, listing techniques in order of increasing possibility of herbicide exposure to the environment or to humans in the vicinity of treated plants.

  1. Stem injection: Herbicides are injected into wounds or cuts in the stems or trunks of plants to be killed. The herbicide must penetrate to the cambial tissue and be water-soluble to be effective. The chemical is then translocated throughout the plant and can provide good root kill, which is important in order to prevent resprouting.
  2. Cut stump treatment: Herbicides are directly applied to the cambial area around the edges of freshly cut stumps. Application must occur within 5-20 minutes of cutting to ensure effectiveness. McHenry (1985) [35] suggests late spring as the best season to do this. In early spring sap may flow to the surface of the cut and rinse the chemical off. At other times of the year translocation is too poor to adequately distribute the chemical. Applications may be made with backpack sprayers, sprinkling cans, brush and pail, or squeeze bottles. Picloram should not be used for this technique as it is known to “flashback” through root grafts between treated and untreated plants and may damage the untreated individuals.
  3. Basal/Stem sprays: High concentrations of herbicides in oil or other penetrating carriers are applied, using backpack sprayers, to the basal portion of stems to be killed. The oil carrier is necessary for the mixture to penetrate bark and enter the vascular system. This method gives good root kill, especially in the fall when vascular fluids are moving toward the roots. This method may be easier to use with small diameter stems than the two previous techniques.
  4. Herbicide pellets: Pelletized or granular herbicides are scattered at the bases of unwanted plants. Subsequent rainfall dissolves the pellets and leaches the herbicide down to the root system. Picloram granules are most effective in the winter and spring [31] and should be applied prior to the end of the winter rains so as to allow the herbicide to penetrate to the root zone. Premature application, with a great deal of rain yet to fall, may leach the picloram below the root zone. Mowing of the top growth prior to the application of granular picloram improves control effectiveness. The presence of litter may retard the infiltration of the chemical into the soil. [31] Because picloram persists in the soil, is highly toxic to other plants and is costly, it is most suitable for small infestations. [36]

Rubus discolor is difficult to control due to its variety of reproductive tactics. It may reproduce by seed, rooting at cane apices, suckering of lateral roots, and from pieces of roots and canes. It becomes established in disturbed and subsequently neglected areas.

Himalaya-berry rarely invades undisturbed sites because seedlings are easily outcompeted by other plants. Rooting at cane apices may be prevented by grazing, cultivation or herbicides. Cultivation or herbicides may be the only effective means of removing the root system, which must be killed to prevent regrowth.

Seed phrase

This page contains sample addresses and/or private keys. Do not send bitcoins to or import any sample keys; you will lose your money.

A seed phrase, seed recovery phrase or backup seed phrase is a list of words which store all the information needed to recover Bitcoin funds on-chain. Wallet software will typically generate a seed phrase and instruct the user to write it down on paper. If the user’s computer breaks or their hard drive becomes corrupted, they can download the same wallet software again and use the paper backup to get their bitcoins back.

Anybody else who discovers the phrase can steal the bitcoins, so it must be kept safe like jewels or cash. For example, it must not be typed into any website.

See also  herijuana seeds

Seed phrases are an excellent way of backing up and storing bitcoins, so they are used by almost all well-regarded wallets. [1]

Seed phrases can only backups funds on the block chain. They cannot store funds involved in off-chain transactions such as Lightning Network or Blinded bearer certificates. Although these technologies are in their infancy as of 2019 so its possible in future seed phrases could be used to backup them.


BIP39 and its flaws

BIP39 is the most common standard used for seed phrases. One notable example is Electrum wallet, which is using its own standard, and for good reasons. BIP39 has some flaws, known in the technical community but not known much wider. They are described here on this electrum doc page. Most seriously, BIP39 flaws mean it is not true to say that backing up a BIP39 seed phrase and name of wallet software is the only thing a user needs to do to keep their money safe. BIP39 works this way because its designers wanted their hardware wallet to also support altcoins. is an attempt at helping with this issue, but ideally there will be a better solution in the future.


An example of a non-BIP39 seed phrase is:

The word order is important.


A simplified explanation of how seed phrases work is that the wallet software has a list of words taken from a dictionary, with each word assigned to a number. The seed phrase can be converted to a number which is used as the seed integer to a deterministic wallet that generates all the key pairs used in the wallet.

The English-language wordlist for the BIP39 standard has 2048 words, so if the phrase contained only 12 random words, the number of possible combinations would be 2048^12 = 2^132 and the phrase would have 132 bits of security. However, some of the data in a BIP39 phrase is not random, [2] so the actual security of a 12-word BIP39 seed phrase is only 128 bits. This is approximately the same strength as all Bitcoin private keys, so most experts consider it to be sufficiently secure. [3]

It is not safe to invent your own seed phrase because humans are bad at generating randomness. The best way is to allow the wallet software to generate a phrase which you write down.

As seed phrases use natural language words, they have excellent error correction. Words written in bad handwriting can often still be read. If one or two letters are missing or unreadable the word can often still be deduced. The word list that the seed phrase words are drawn from is carefully chosen so that the first four letters of each word are enough to uniquely identify it. This compares well with writing down a raw private key where a single letter being unreadable or incorrect can make the private key useless (depending on the serialization format).

Two-factor seed phrases

Seed phrases, like all backups, can store any amount of bitcoins. It’s a concerning idea to possibly have enough money to purchase the entire building just sitting on a sheet of paper without any protection. For this reason many wallets make it possible to encrypt a seed phrase with a password.

The password can be used to create a two-factor seed phrase where both “something you have” plus “something you know” is required to unlock the bitcoins.

This works by the wallet creating a seed phrase and asking the user for a password. Then both the seed phrase and extra word are required to recover the wallet. Electrum and some other wallets call the passphrase a “seed extension”, “extension word” or “13th/25th word”. The BIP39 standard defines a way of passphrase-protecting a seed phrase. A similar scheme is also used in the Electrum standard. If a passphrase is not present, an empty string “” is used instead.

Warning: Forgetting this password will result in the bitcoin wallet and any contained money being lost. Do not overestimate your ability to remember passphrases especially when you may not use it very often.

Warning: The seed phrase password should not be confused with the password used to encrypt the wallet file on disk. This is probably why many wallets call it an extension word instead of a password.

Storing seed phrases for the long term

Most people write down phrases on paper but they can be stored in many other ways such as memorizing, engraving or stamping on metal, writing in the margins of a book, chiselling into a stone tablet or any other creative and inventive way.

In the past many people have accidentally lost bitcoins because of failed backups, mistyped letters, forgotten hard drives, or corrupted SSD devices. It’s also important to protect the seed from accidental loss.

It could be a good idea to write some words of explanation on the same paper as the seed phrase. If storing for the long term you may forget what a phrase is how it should be treated. A sample explanation that can be adapted is:

Paper and pencil backup

Through bitter experience it has been found that one of the most practical storage mediums is pencil and paper. The private keys of a bitcoin wallet are encoded into random words from a dictionary which can be written down. If your hard drive crashes, you can find the paper with the seed phrase and restore the entire wallet. As seed phrases use natural language words, they have good error correction. Words written in bad handwriting can often still be read. If one or two letters are missing the word can often still be deduced. The word list that the seed phrase words are drawn from is carefully chosen so that the first four letters of each word are enough to uniquely identify it.

For storing on paper writing with pencil is much better than pen [4] [5] . Paper should be acid-free or archival paper, and stored in the dark avoiding extremes of heat and moisture [6] [7] [8] .

Metal backup

Seed phrases can also be stamped or engraved into metal which is significantly more durable than paper. Metal backups are recommended if the threat model involves fire, water, extremes of temperature or physical stress.

Methods that are not recommended

Some methods that are not recommended are: storing in a file on a computer (including online), or storing online.

Some people get the idea to split up their phrases, like storing 6 words in one location and the other 6 words in another location. This is a bad idea and should not be done, because if one set of 6 words is discovered then it becomes far easier to brute-force the rest of the phrase. Storing bitcoins in multiple locations like this should be done with multi-signature wallets instead.

The Shamir Secret Sharing algorithm is sometimes promoted as a way to divide control of bitcoins, but in practice there are many pitfalls and trade-offs that make it not worth it. [9]

Another bad idea is to add random decoy words that are somehow meaningful to you and later remove them to be left with only the 12-word phrase. The phrase words come from a known dictionary (see next section), so anybody can use that dictionary to weed out the decoy words.

It’s possible but risky to memorize (Brainwallets) seed phrases. This should probably only be done in situations that really need it, such as crossing a hostile border where one expects to be searched.

Word lists

Generally, a seed phrase only works with the same wallet software that created it. If storing for a long period of time it’s a good idea to write the name of the wallet too.

The BIP39 English word list has each word being uniquely identified by the first four letters, which can be useful when space to write them is scarce.

  • BIP39 wordlists
  • Electrum old-style wordlist
  • Electrum new-style wordlist

Alternative name “mnemonic phrase”

Seed phrases are sometimes called mnemonic phrases, especially in older literature. This is a bad name because the word “mnemonic” implies that the phrase should be memorized. It is less misleading to call them seed phrases.

The power of backups

An especially interesting aspect in the power of paper backups is allowing your money to be two places at once. At the London Inside Bitcoin conference, the keynote speaker showed 25 paper backups they were carrying—all password-protected. With that, one can carry $100,000 which can instantly be moved to a phone or transferred yet with total security. If it’s stolen, then there is no risk because it is backed up elsewhere. That is powerful. [10]